HealthCare_ADM_with_DL_updated.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "HealthCare-ADM-with-DL-updated.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyPS4vVoz4ALhY9qni6gYEhh",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/MuhammadShifa/Anamoly-Detection-Model-ADM-in-Health-Care-Insurance-/blob/main/HealthCare_ADM_with_DL_updated.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h2xpCd3jn5Ko",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dec6f9a1-94dd-464e-8b8b-9f41063e5f40"
      },
      "source": [
        "# To access the drive data\r\n",
        "from google.colab import drive\r\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Hy6C-r0JetlR"
      },
      "source": [
        "## Load the Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 396
        },
        "id": "y6L6cayZysTU",
        "outputId": "0187b074-c010-48aa-8af7-1e5481170f4d"
      },
      "source": [
        "import pandas as pd\r\n",
        "\r\n",
        "# Load the dataset to pandas dataframe\r\n",
        "dataframe = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "print(dataframe.shape)\r\n",
        "\r\n",
        "# Print first 10 value of data frame\r\n",
        "dataframe.head(10)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(558211, 27)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Provider</th>\n",
              "      <th>PotentialFraud</th>\n",
              "      <th>InscClaimAmtReimbursed</th>\n",
              "      <th>DeductibleAmtPaid</th>\n",
              "      <th>AdmitDays</th>\n",
              "      <th>Gender</th>\n",
              "      <th>Race</th>\n",
              "      <th>RenalDiseaseIndicator</th>\n",
              "      <th>NoOfMonths_PartACov</th>\n",
              "      <th>NoOfMonths_PartBCov</th>\n",
              "      <th>ChronicCond_Alzheimer</th>\n",
              "      <th>ChronicCond_Heartfailure</th>\n",
              "      <th>ChronicCond_KidneyDisease</th>\n",
              "      <th>ChronicCond_Cancer</th>\n",
              "      <th>ChronicCond_ObstrPulmonary</th>\n",
              "      <th>ChronicCond_Depression</th>\n",
              "      <th>ChronicCond_Diabetes</th>\n",
              "      <th>ChronicCond_IschemicHeart</th>\n",
              "      <th>ChronicCond_Osteoporasis</th>\n",
              "      <th>ChronicCond_rheumatoidarthritis</th>\n",
              "      <th>ChronicCond_stroke</th>\n",
              "      <th>IPAnnualReimbursementAmt</th>\n",
              "      <th>IPAnnualDeductibleAmt</th>\n",
              "      <th>OPAnnualReimbursementAmt</th>\n",
              "      <th>OPAnnualDeductibleAmt</th>\n",
              "      <th>Age</th>\n",
              "      <th>IsDead</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>20</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>300</td>\n",
              "      <td>110</td>\n",
              "      <td>80.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>700</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>720</td>\n",
              "      <td>10</td>\n",
              "      <td>67.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>900</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1380</td>\n",
              "      <td>370</td>\n",
              "      <td>76.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>500</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2020</td>\n",
              "      <td>1068</td>\n",
              "      <td>6700</td>\n",
              "      <td>2700</td>\n",
              "      <td>74.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>36000</td>\n",
              "      <td>1068.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>36000</td>\n",
              "      <td>1068</td>\n",
              "      <td>3520</td>\n",
              "      <td>140</td>\n",
              "      <td>69.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>12000</td>\n",
              "      <td>1068.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>95000</td>\n",
              "      <td>2136</td>\n",
              "      <td>2270</td>\n",
              "      <td>900</td>\n",
              "      <td>72.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>70</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>5000</td>\n",
              "      <td>1068</td>\n",
              "      <td>370</td>\n",
              "      <td>70</td>\n",
              "      <td>91.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>100</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>320</td>\n",
              "      <td>80</td>\n",
              "      <td>90.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>300</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>630</td>\n",
              "      <td>130</td>\n",
              "      <td>57.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>PRV51001</td>\n",
              "      <td>0</td>\n",
              "      <td>60</td>\n",
              "      <td>0.0</td>\n",
              "      <td>7.0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "      <td>12</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1590</td>\n",
              "      <td>1070</td>\n",
              "      <td>98.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Provider  PotentialFraud  ...   Age  IsDead\n",
              "0  PRV51001               0  ...  80.0     0.0\n",
              "1  PRV51001               0  ...  67.0     0.0\n",
              "2  PRV51001               0  ...  76.0     0.0\n",
              "3  PRV51001               0  ...  74.0     0.0\n",
              "4  PRV51001               0  ...  69.0     0.0\n",
              "5  PRV51001               0  ...  72.0     0.0\n",
              "6  PRV51001               0  ...  91.0     0.0\n",
              "7  PRV51001               0  ...  90.0     0.0\n",
              "8  PRV51001               0  ...  57.0     0.0\n",
              "9  PRV51001               0  ...  98.0     0.0\n",
              "\n",
              "[10 rows x 27 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gbbWGv7BfFcz"
      },
      "source": [
        "## Cleaning the Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M_CohuebqBo2",
        "outputId": "e894d027-fb21-494d-c67b-9c4d5a9f465e"
      },
      "source": [
        "# print data information i,e no of columns and rows etc\r\n",
        "data_info = dataframe.info()\r\n",
        "print(data_info)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 558211 entries, 0 to 558210\n",
            "Data columns (total 27 columns):\n",
            " #   Column                           Non-Null Count   Dtype  \n",
            "---  ------                           --------------   -----  \n",
            " 0   Provider                         558211 non-null  object \n",
            " 1   PotentialFraud                   558211 non-null  int64  \n",
            " 2   InscClaimAmtReimbursed           558211 non-null  int64  \n",
            " 3   DeductibleAmtPaid                558211 non-null  float64\n",
            " 4   AdmitDays                        558211 non-null  float64\n",
            " 5   Gender                           558211 non-null  int64  \n",
            " 6   Race                             558211 non-null  int64  \n",
            " 7   RenalDiseaseIndicator            558211 non-null  int64  \n",
            " 8   NoOfMonths_PartACov              558211 non-null  int64  \n",
            " 9   NoOfMonths_PartBCov              558211 non-null  int64  \n",
            " 10  ChronicCond_Alzheimer            558211 non-null  int64  \n",
            " 11  ChronicCond_Heartfailure         558211 non-null  int64  \n",
            " 12  ChronicCond_KidneyDisease        558211 non-null  int64  \n",
            " 13  ChronicCond_Cancer               558211 non-null  int64  \n",
            " 14  ChronicCond_ObstrPulmonary       558211 non-null  int64  \n",
            " 15  ChronicCond_Depression           558211 non-null  int64  \n",
            " 16  ChronicCond_Diabetes             558211 non-null  int64  \n",
            " 17  ChronicCond_IschemicHeart        558211 non-null  int64  \n",
            " 18  ChronicCond_Osteoporasis         558211 non-null  int64  \n",
            " 19  ChronicCond_rheumatoidarthritis  558211 non-null  int64  \n",
            " 20  ChronicCond_stroke               558211 non-null  int64  \n",
            " 21  IPAnnualReimbursementAmt         558211 non-null  int64  \n",
            " 22  IPAnnualDeductibleAmt            558211 non-null  int64  \n",
            " 23  OPAnnualReimbursementAmt         558211 non-null  int64  \n",
            " 24  OPAnnualDeductibleAmt            558211 non-null  int64  \n",
            " 25  Age                              558211 non-null  float64\n",
            " 26  IsDead                           558211 non-null  float64\n",
            "dtypes: float64(4), int64(22), object(1)\n",
            "memory usage: 115.0+ MB\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uwKbsoVap7SY",
        "outputId": "62cf9894-52a1-4850-f467-a025ef9b4ae7"
      },
      "source": [
        "# Check is there any null value\r\n",
        "dataframe.isnull().sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Provider                           0\n",
              "PotentialFraud                     0\n",
              "InscClaimAmtReimbursed             0\n",
              "DeductibleAmtPaid                  0\n",
              "AdmitDays                          0\n",
              "Gender                             0\n",
              "Race                               0\n",
              "RenalDiseaseIndicator              0\n",
              "NoOfMonths_PartACov                0\n",
              "NoOfMonths_PartBCov                0\n",
              "ChronicCond_Alzheimer              0\n",
              "ChronicCond_Heartfailure           0\n",
              "ChronicCond_KidneyDisease          0\n",
              "ChronicCond_Cancer                 0\n",
              "ChronicCond_ObstrPulmonary         0\n",
              "ChronicCond_Depression             0\n",
              "ChronicCond_Diabetes               0\n",
              "ChronicCond_IschemicHeart          0\n",
              "ChronicCond_Osteoporasis           0\n",
              "ChronicCond_rheumatoidarthritis    0\n",
              "ChronicCond_stroke                 0\n",
              "IPAnnualReimbursementAmt           0\n",
              "IPAnnualDeductibleAmt              0\n",
              "OPAnnualReimbursementAmt           0\n",
              "OPAnnualDeductibleAmt              0\n",
              "Age                                0\n",
              "IsDead                             0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2bMFGed6iOLl",
        "outputId": "90f08ffd-e91d-433b-ace9-fbceacd5f21a"
      },
      "source": [
        "# Checking is there any duplicate rows\r\n",
        "print(\"The shape of dataframe is: \",dataframe.shape)\r\n",
        "dataframe = dataframe.drop_duplicates(subset = None, )\r\n",
        "print(\"The shape of dataframe after duplicate row: \", dataframe.shape)\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The shape of dataframe is:  (558211, 27)\n",
            "The shape of dataframe after duplicate row:  (534843, 27)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ImXQ9ZxwszZx"
      },
      "source": [
        "As we have found 23, 368 rows duplicate so we will drop this."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UC-a9lc_fUIs"
      },
      "source": [
        "## Exploratory Data Analysis (EDA)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jzoDRSiNhjSi"
      },
      "source": [
        "#### Dependent Variable\r\n",
        "\r\n",
        "Exploratory data analysis was conducted started with the dependent variable, PotentialFraud in our case. There were 201403 frauds and 333440 non-frauds. 37.65% of the data were frauds while 62.34% were non-fraudulent claims. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Dj2CychyrFCo",
        "outputId": "2e00c663-02d7-4549-eb1a-1755472149d5"
      },
      "source": [
        "# Count the occurence of Fraud/non-fraud \r\n",
        "occ = dataframe['PotentialFraud'].value_counts()\r\n",
        "print(occ)\r\n",
        "\r\n",
        "# Length of total row/ observation\r\n",
        "total_rows =len(dataframe.index)\r\n",
        "print(total_rows)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0    333440\n",
            "1    201403\n",
            "Name: PotentialFraud, dtype: int64\n",
            "534843\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zNm9MqaCiZqB",
        "outputId": "9afbe261-520f-462b-cf1f-fa446483c557"
      },
      "source": [
        "occ_0, occ_1 = occ\r\n",
        "print(\"The occ of 0 is: \",occ_0)\r\n",
        "print(\"The occ of 1 is: \",occ_1)\r\n",
        "\r\n",
        "# Lets check the percentage of 0 and 1 in our dataset\r\n",
        "\r\n",
        "percent_0 = (occ_0/total_rows) * 100\r\n",
        "percent_1 = (occ_1/total_rows) * 100\r\n",
        "\r\n",
        "print(\"The Percentage of 0 in dataset:\",percent_0)\r\n",
        "print(\"The Percentage of 1 in dataset\", percent_1)\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The occ of 0 is:  333440\n",
            "The occ of 1 is:  201403\n",
            "The Percentage of 0 in dataset: 62.34352884865279\n",
            "The Percentage of 1 in dataset 37.65647115134721\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4kPyXdmoreno",
        "outputId": "19f99230-03a7-4b3b-b50d-85fc703a707d"
      },
      "source": [
        "# For checking Imbalanced classification let's check the ratio of fraud and non-fraud\r\n",
        "ratio_cases = occ/(len(dataframe.index))\r\n",
        "print(f'Ratio of non-farudulent cases: {ratio_cases[0]} \\nRatio of fraudulent case: {ratio_cases[1]}')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Ratio of non-farudulent cases: 0.6234352884865278 \n",
            "Ratio of fraudulent case: 0.37656471151347215\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3lqq3PLifgx-"
      },
      "source": [
        "The above result of fradulent and non-fraudulent ratio shows that we an consider it as a balanced classification upto some extent. As we have not a lot imabalanced classifcation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JblyL1FJlwWW"
      },
      "source": [
        "#### Data Visualization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "9Y9iyeSDvlZe",
        "outputId": "d963e779-5b1c-4762-c415-e70ac5a2b78d"
      },
      "source": [
        "# visualize the target variable i,e PotentialFraud here\r\n",
        "import seaborn as sns\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "g = sns.countplot(dataframe['PotentialFraud'])\r\n",
        "g.set_xticklabels(['Not Fraud','Fraud'])\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/seaborn/_decorators.py:43: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
            "  FutureWarning\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8oTLXlWL2H1o"
      },
      "source": [
        "import numpy as np\r\n",
        "# Convert the DataFrame into two numpy variable\r\n",
        "def prep_data(df: pd.DataFrame) -> (np.ndarray, np.ndarray):\r\n",
        "    X = df.iloc[:, 2:28].values\r\n",
        "    y = df.PotentialFraud.values\r\n",
        "\r\n",
        "    return X, y\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CZEyCV-k201r"
      },
      "source": [
        "%matplotlib inline\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "# Define a function to create a scatter plot of our data and labels\r\n",
        "def plot_data(X: np.ndarray, y: np.ndarray):\r\n",
        "    plt.scatter(X[y == 0, 0], X[y == 0, 1], label=\"Class #0\", alpha=0.5, linewidth=0.15)\r\n",
        "    plt.scatter(X[y == 1, 0], X[y == 1, 1], label=\"Class #1\", alpha=0.5, linewidth=0.15, c='r')\r\n",
        "    plt.legend()\r\n",
        "    return plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m3daxs6x3Vxa",
        "outputId": "4940e03c-1e6f-4a3b-e8f6-550e2d0f6872"
      },
      "source": [
        "# Create X and y from the prep_data function \r\n",
        "X, y = prep_data(dataframe)\r\n",
        "print(X.shape)\r\n",
        "print(y.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(534843, 25)\n",
            "(534843,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "pmdU7Mk43Vur",
        "outputId": "f0d7d07c-06e4-485b-9a9d-b4e37338440a"
      },
      "source": [
        "# Plot our data by running our plot data function on X and y\r\n",
        "plot_data(X, y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "KqSnqZZf3Vr1",
        "outputId": "1efd3697-5f54-47e2-c9ca-c334c6c06627"
      },
      "source": [
        "plt.scatter(dataframe.InscClaimAmtReimbursed[dataframe.PotentialFraud == 0], dataframe.DeductibleAmtPaid[dataframe.PotentialFraud == 0], label=\"Class #0\", alpha=0.5, linewidth=0.15)\r\n",
        "plt.scatter(dataframe.InscClaimAmtReimbursed[dataframe.PotentialFraud == 1], dataframe.DeductibleAmtPaid[dataframe.PotentialFraud == 1], label=\"Class #1\", alpha=0.5, linewidth=0.15, c='r')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vHjK2ivK4jOK",
        "outputId": "268b669e-3f61-4043-8661-019d63c90ca4"
      },
      "source": [
        "# Converting pandas dataframe to x and y\r\n",
        "X = dataframe.drop(['Provider','PotentialFraud'], axis = 1).astype(float)\r\n",
        "y = dataframe.PotentialFraud.astype(float)\r\n",
        "print(X)\r\n",
        "print(y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "        InscClaimAmtReimbursed  DeductibleAmtPaid  ...   Age  IsDead\n",
            "0                         20.0                0.0  ...  80.0     0.0\n",
            "1                        700.0                0.0  ...  67.0     0.0\n",
            "2                        900.0                0.0  ...  76.0     0.0\n",
            "3                        500.0                0.0  ...  74.0     0.0\n",
            "4                      36000.0             1068.0  ...  69.0     0.0\n",
            "...                        ...                ...  ...   ...     ...\n",
            "558206                  1100.0               40.0  ...  70.0     0.0\n",
            "558207                   200.0                0.0  ...  51.0     0.0\n",
            "558208                    80.0                0.0  ...  46.0     0.0\n",
            "558209                  3300.0                0.0  ...  58.0     0.0\n",
            "558210                    40.0                0.0  ...  58.0     0.0\n",
            "\n",
            "[558211 rows x 25 columns]\n",
            "0         0.0\n",
            "1         0.0\n",
            "2         0.0\n",
            "3         0.0\n",
            "4         0.0\n",
            "         ... \n",
            "558206    0.0\n",
            "558207    0.0\n",
            "558208    0.0\n",
            "558209    0.0\n",
            "558210    0.0\n",
            "Name: PotentialFraud, Length: 558211, dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xs2_8LG67RYu",
        "outputId": "06752744-df3a-4a53-ee67-4329b83c4d2b"
      },
      "source": [
        "# find the variance and correlation in the features\r\n",
        "print(X.var())\r\n",
        "print(X.corr())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "InscClaimAmtReimbursed             1.519090e+07\n",
            "DeductibleAmtPaid                  7.785084e+04\n",
            "AdmitDays                          2.408837e+00\n",
            "Gender                             2.438275e-01\n",
            "Race                               5.155307e-01\n",
            "RenalDiseaseIndicator              1.574123e-01\n",
            "NoOfMonths_PartACov                8.031344e-01\n",
            "NoOfMonths_PartBCov                6.252241e-01\n",
            "ChronicCond_Alzheimer              2.400265e-01\n",
            "ChronicCond_Heartfailure           2.423371e-01\n",
            "ChronicCond_KidneyDisease          2.418730e-01\n",
            "ChronicCond_Cancer                 1.279403e-01\n",
            "ChronicCond_ObstrPulmonary         2.144797e-01\n",
            "ChronicCond_Depression             2.454552e-01\n",
            "ChronicCond_Diabetes               2.089413e-01\n",
            "ChronicCond_IschemicHeart          1.839640e-01\n",
            "ChronicCond_Osteoporasis           2.163934e-01\n",
            "ChronicCond_rheumatoidarthritis    2.138355e-01\n",
            "ChronicCond_stroke                 9.110468e-02\n",
            "IPAnnualReimbursementAmt           1.398836e+08\n",
            "IPAnnualDeductibleAmt              1.392688e+06\n",
            "OPAnnualReimbursementAmt           1.492203e+07\n",
            "OPAnnualDeductibleAmt              9.924359e+05\n",
            "Age                                1.694228e+02\n",
            "IsDead                             7.456061e-03\n",
            "dtype: float64\n",
            "                                 InscClaimAmtReimbursed  ...    IsDead\n",
            "InscClaimAmtReimbursed                         1.000000  ...  0.001091\n",
            "DeductibleAmtPaid                              0.654597  ...  0.001066\n",
            "AdmitDays                                      0.195705  ...  0.000078\n",
            "Gender                                        -0.001975  ...  0.001592\n",
            "Race                                          -0.000678  ...  0.003289\n",
            "RenalDiseaseIndicator                          0.061018  ... -0.004169\n",
            "NoOfMonths_PartACov                           -0.005356  ...  0.000805\n",
            "NoOfMonths_PartBCov                           -0.005716  ...  0.001728\n",
            "ChronicCond_Alzheimer                         -0.035506  ... -0.009117\n",
            "ChronicCond_Heartfailure                      -0.060829  ... -0.003222\n",
            "ChronicCond_KidneyDisease                     -0.099764  ... -0.000054\n",
            "ChronicCond_Cancer                            -0.027640  ... -0.003982\n",
            "ChronicCond_ObstrPulmonary                    -0.070623  ... -0.002958\n",
            "ChronicCond_Depression                        -0.019326  ...  0.003720\n",
            "ChronicCond_Diabetes                          -0.046858  ... -0.005155\n",
            "ChronicCond_IschemicHeart                     -0.054386  ... -0.007748\n",
            "ChronicCond_Osteoporasis                      -0.011142  ...  0.001602\n",
            "ChronicCond_rheumatoidarthritis               -0.018469  ... -0.001082\n",
            "ChronicCond_stroke                            -0.049893  ...  0.003899\n",
            "IPAnnualReimbursementAmt                       0.390002  ... -0.007418\n",
            "IPAnnualDeductibleAmt                          0.229470  ... -0.005954\n",
            "OPAnnualReimbursementAmt                       0.063597  ... -0.008685\n",
            "OPAnnualDeductibleAmt                          0.030587  ... -0.011508\n",
            "Age                                            0.005799  ...  0.006217\n",
            "IsDead                                         0.001091  ...  1.000000\n",
            "\n",
            "[25 rows x 25 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "-02M-VvMzabk",
        "outputId": "367581c4-4bd8-40cd-bdd6-a08be35cba6f"
      },
      "source": [
        "import numpy as np\r\n",
        "import seaborn as sn\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "corrMatt = dataframe[['PotentialFraud','DeductibleAmtPaid','OPAnnualReimbursementAmt','OPAnnualDeductibleAmt','Race','Age',\r\n",
        "                    'IPAnnualReimbursementAmt']].corr()\r\n",
        "mask = np.array(corrMatt)\r\n",
        "mask[np.tril_indices_from(mask)] = False\r\n",
        "fig,ax= plt.subplots()\r\n",
        "fig.set_size_inches(30,15)\r\n",
        "sn.heatmap(corrMatt, mask=mask,vmax=.8, square=True,annot=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f591b475c88>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x1080 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "EUPQUXYf7ROE",
        "outputId": "9f540210-8528-4e20-a3fd-b36f85f1a032"
      },
      "source": [
        "# Check the relative importance of Features with RandomForestRegressor\r\n",
        "import numpy\r\n",
        "from sklearn.decomposition import PCA\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "from sklearn.ensemble import RandomForestRegressor\r\n",
        "\r\n",
        "model = RandomForestRegressor(random_state=1, max_depth=10)\r\n",
        "df=pd.get_dummies(X)\r\n",
        "model.fit(df,y)\r\n",
        "import numpy as np\r\n",
        "features = X.columns\r\n",
        "importances = model.feature_importances_\r\n",
        "indices = np.argsort(importances)[-23:]  \r\n",
        "plt.title('Feature Importances')\r\n",
        "plt.barh(range(len(indices)), importances[indices], color='b', align='center')\r\n",
        "plt.yticks(range(len(indices)), [features[i] for i in indices])\r\n",
        "plt.xlabel('Relative Importance')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Fpvgf3YQ9kzq",
        "outputId": "8daeda27-a7f0-460e-a85f-5d0ddd6e15f5"
      },
      "source": [
        "#Standardizing the data with StandardScaler mean = 0 and std = 1\r\n",
        "from sklearn.preprocessing import StandardScaler\r\n",
        "#dataset import\r\n",
        "\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = dataframe.iloc[:,2:27].values\r\n",
        "y = dataframe.iloc[:,1].values\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "sc = StandardScaler()\r\n",
        "X = sc.fit_transform(X)\r\n",
        "print(X)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[-0.25985153 -0.29290773  0.01628602 ... -0.53380379  0.47878796\n",
            "  -0.08700201]\n",
            " [-0.08538284 -0.29290773  0.01628602 ... -0.63418425 -0.51996429\n",
            "  -0.08700201]\n",
            " [-0.03406852 -0.29290773  0.01628602 ... -0.27281461  0.17147957\n",
            "  -0.08700201]\n",
            " ...\n",
            " [-0.24445723 -0.29290773  0.01628602 ... -0.14232001 -2.13333329\n",
            "  -0.08700201]\n",
            " [ 0.58170333 -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]\n",
            " [-0.2547201  -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tnWOjSJG7RFr",
        "outputId": "e8b638b2-d336-4f56-c7d8-c84132dae37e"
      },
      "source": [
        "#Scaling the data with MinMaxScaler\r\n",
        "from sklearn.preprocessing import MinMaxScaler\r\n",
        "\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = dataframe.iloc[:,2:27].values\r\n",
        "y = dataframe.iloc[:,1:2].values\r\n",
        "\r\n",
        "\r\n",
        "scaler = MinMaxScaler(feature_range=(0,1))\r\n",
        "X = scaler.fit_transform(X)\r\n",
        "print(X)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1.60000000e-04 0.00000000e+00 1.71428571e-01 ... 7.94797688e-03\n",
            "  7.20000000e-01 0.00000000e+00]\n",
            " [5.60000000e-03 0.00000000e+00 1.71428571e-01 ... 7.22543353e-04\n",
            "  5.46666667e-01 0.00000000e+00]\n",
            " [7.20000000e-03 0.00000000e+00 1.71428571e-01 ... 2.67341040e-02\n",
            "  6.66666667e-01 0.00000000e+00]\n",
            " ...\n",
            " [6.40000000e-04 0.00000000e+00 1.71428571e-01 ... 3.61271676e-02\n",
            "  2.66666667e-01 0.00000000e+00]\n",
            " [2.64000000e-02 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]\n",
            " [3.20000000e-04 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DeWj-Zqz_aPs",
        "outputId": "704aa03a-b4e2-44ad-b01c-1c25b3a2b664"
      },
      "source": [
        "# converting the output class to onehotencoding 'Which is the process of converting the output class to binary value\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "y"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       ...,\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tb1_q1YjByFI"
      },
      "source": [
        "## Modeling or Keras-Classifier\r\n",
        "\r\n",
        "1. Multi-Layer Perceptron (MLP)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sLT-sANCB6FG",
        "outputId": "150c6e62-c48a-446e-a842-06d9ba42ecd9"
      },
      "source": [
        "# 1. Simple MLP for classification with all features without normazlizing and standardizing\r\n",
        "import pandas as pd\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "import keras\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense,BatchNormalization,Dropout\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "#Changing pandas dataset to numpy array\r\n",
        "X = dataset.iloc[:,2:27].values\r\n",
        "y = dataset.iloc[:,1].values\r\n",
        "\r\n",
        "#Spliting into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.25)\r\n",
        "print(X_train.shape)\r\n",
        "print(y_train.shape)\r\n",
        "\r\n",
        "# Neural network\r\n",
        "model = Sequential()\r\n",
        "model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.25))\r\n",
        "model.add(Dense(1, activation='sigmoid'))\r\n",
        "# compile the keras model\r\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=20, batch_size=32, validation_data=(X_test, y_test),  verbose=1)\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(401132, 25)\n",
            "(401132,)\n",
            "Epoch 1/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6853 - accuracy: 0.6156 - val_loss: 0.6581 - val_accuracy: 0.6339\n",
            "Epoch 2/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6572 - accuracy: 0.6318 - val_loss: 0.6569 - val_accuracy: 0.6317\n",
            "Epoch 3/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6568 - accuracy: 0.6327 - val_loss: 0.6567 - val_accuracy: 0.6339\n",
            "Epoch 4/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6573 - accuracy: 0.6316 - val_loss: 0.6564 - val_accuracy: 0.6339\n",
            "Epoch 5/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6568 - accuracy: 0.6324 - val_loss: 0.6576 - val_accuracy: 0.6334\n",
            "Epoch 6/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6572 - accuracy: 0.6315 - val_loss: 0.6574 - val_accuracy: 0.6337\n",
            "Epoch 7/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6565 - accuracy: 0.6330 - val_loss: 0.6566 - val_accuracy: 0.6331\n",
            "Epoch 8/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6559 - accuracy: 0.6343 - val_loss: 0.6573 - val_accuracy: 0.6321\n",
            "Epoch 9/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6569 - accuracy: 0.6327 - val_loss: 0.6565 - val_accuracy: 0.6334\n",
            "Epoch 10/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6564 - accuracy: 0.6333 - val_loss: 0.6566 - val_accuracy: 0.6337\n",
            "Epoch 11/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6569 - accuracy: 0.6321 - val_loss: 0.6565 - val_accuracy: 0.6339\n",
            "Epoch 12/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6565 - accuracy: 0.6331 - val_loss: 0.6565 - val_accuracy: 0.6339\n",
            "Epoch 13/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6564 - accuracy: 0.6338 - val_loss: 0.6566 - val_accuracy: 0.6339\n",
            "Epoch 14/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6561 - accuracy: 0.6335 - val_loss: 0.6563 - val_accuracy: 0.6337\n",
            "Epoch 15/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6561 - accuracy: 0.6331 - val_loss: 0.6567 - val_accuracy: 0.6339\n",
            "Epoch 16/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6559 - accuracy: 0.6342 - val_loss: 0.6565 - val_accuracy: 0.6338\n",
            "Epoch 17/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6566 - accuracy: 0.6332 - val_loss: 0.6564 - val_accuracy: 0.6339\n",
            "Epoch 18/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6558 - accuracy: 0.6349 - val_loss: 0.6564 - val_accuracy: 0.6339\n",
            "Epoch 19/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6570 - accuracy: 0.6318 - val_loss: 0.6566 - val_accuracy: 0.6338\n",
            "Epoch 20/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6571 - accuracy: 0.6318 - val_loss: 0.6567 - val_accuracy: 0.6339\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Wepeans-FF6D",
        "outputId": "d9019ed1-dfcb-48b2-f362-c9573daac17b"
      },
      "source": [
        "# Simple MLP for classification with Dimensionality Reduction by droping some features\r\n",
        "import pandas as pd\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "\r\n",
        "#Spliting into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.25)\r\n",
        "print(X_train.shape)\r\n",
        "print(y_train.shape)\r\n",
        "\r\n",
        "# Neural network\r\n",
        "model = Sequential()\r\n",
        "model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.25))\r\n",
        "model.add(Dense(1, activation='sigmoid'))\r\n",
        "# compile the keras model\r\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=20, batch_size=32, validation_data=(X_test, y_test),  verbose=1)\r\n",
        "\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(401132, 14)\n",
            "(401132,)\n",
            "Epoch 1/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6741 - accuracy: 0.6185 - val_loss: 0.6573 - val_accuracy: 0.6294\n",
            "Epoch 2/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6570 - accuracy: 0.6321 - val_loss: 0.6560 - val_accuracy: 0.6344\n",
            "Epoch 3/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6567 - accuracy: 0.6323 - val_loss: 0.6582 - val_accuracy: 0.6349\n",
            "Epoch 4/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6560 - accuracy: 0.6343 - val_loss: 0.6566 - val_accuracy: 0.6309\n",
            "Epoch 5/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6572 - accuracy: 0.6314 - val_loss: 0.6567 - val_accuracy: 0.6346\n",
            "Epoch 6/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6565 - accuracy: 0.6326 - val_loss: 0.6558 - val_accuracy: 0.6349\n",
            "Epoch 7/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6572 - accuracy: 0.6320 - val_loss: 0.6556 - val_accuracy: 0.6350\n",
            "Epoch 8/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6565 - accuracy: 0.6324 - val_loss: 0.6557 - val_accuracy: 0.6350\n",
            "Epoch 9/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6568 - accuracy: 0.6326 - val_loss: 0.6559 - val_accuracy: 0.6349\n",
            "Epoch 10/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6559 - accuracy: 0.6336 - val_loss: 0.6557 - val_accuracy: 0.6351\n",
            "Epoch 11/20\n",
            "12536/12536 [==============================] - 19s 2ms/step - loss: 0.6569 - accuracy: 0.6323 - val_loss: 0.6865 - val_accuracy: 0.6351\n",
            "Epoch 12/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6567 - accuracy: 0.6322 - val_loss: 0.6562 - val_accuracy: 0.6351\n",
            "Epoch 13/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6558 - accuracy: 0.6341 - val_loss: 0.6560 - val_accuracy: 0.6351\n",
            "Epoch 14/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6567 - accuracy: 0.6326 - val_loss: 0.6557 - val_accuracy: 0.6350\n",
            "Epoch 15/20\n",
            "12536/12536 [==============================] - 21s 2ms/step - loss: 0.6569 - accuracy: 0.6328 - val_loss: 0.6556 - val_accuracy: 0.6352\n",
            "Epoch 16/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6566 - accuracy: 0.6336 - val_loss: 0.6557 - val_accuracy: 0.6352\n",
            "Epoch 17/20\n",
            "12536/12536 [==============================] - 19s 2ms/step - loss: 0.6573 - accuracy: 0.6318 - val_loss: 0.6560 - val_accuracy: 0.6344\n",
            "Epoch 18/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6570 - accuracy: 0.6316 - val_loss: 0.6557 - val_accuracy: 0.6351\n",
            "Epoch 19/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6562 - accuracy: 0.6342 - val_loss: 0.6561 - val_accuracy: 0.6352\n",
            "Epoch 20/20\n",
            "12536/12536 [==============================] - 20s 2ms/step - loss: 0.6559 - accuracy: 0.6339 - val_loss: 0.6555 - val_accuracy: 0.6352\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cKXxoopvKoiH",
        "outputId": "dfaeb11f-2a77-48cf-85a7-2499a3eed3ce"
      },
      "source": [
        "# Simple MLP for classification with Dimensionality resduction and scaling input\r\n",
        "import pandas as pd\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense,BatchNormalization, Dropout\r\n",
        "from sklearn.preprocessing import MinMaxScaler\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "y = y.reshape(-1,1)\r\n",
        "\r\n",
        "# Scaling the data with MinMaxScaler\r\n",
        "\r\n",
        "scaler =MinMaxScaler(feature_range=(0,1))\r\n",
        "scaledX = scaler.fit_transform(X)\r\n",
        "print(scaledX)\r\n",
        "\r\n",
        "# one hot encoding the y\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "\r\n",
        "#Split into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(scaledX,y,test_size = 0.25)\r\n",
        "print(X_train.shape)\r\n",
        "print(y_train.shape)\r\n",
        "\r\n",
        "# Neural network\r\n",
        "model = Sequential()\r\n",
        "model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.25))\r\n",
        "model.add(Dense(2, activation='sigmoid'))\r\n",
        "# compile the keras model\r\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=20, batch_size=32, validation_data=(X_test, y_test),  verbose=1)\r\n",
        "\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1.60000000e-04 0.00000000e+00 1.71428571e-01 ... 7.94797688e-03\n",
            "  7.20000000e-01 0.00000000e+00]\n",
            " [5.60000000e-03 0.00000000e+00 1.71428571e-01 ... 7.22543353e-04\n",
            "  5.46666667e-01 0.00000000e+00]\n",
            " [7.20000000e-03 0.00000000e+00 1.71428571e-01 ... 2.67341040e-02\n",
            "  6.66666667e-01 0.00000000e+00]\n",
            " ...\n",
            " [6.40000000e-04 0.00000000e+00 1.71428571e-01 ... 3.61271676e-02\n",
            "  2.66666667e-01 0.00000000e+00]\n",
            " [2.64000000e-02 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]\n",
            " [3.20000000e-04 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]]\n",
            "(401132, 14)\n",
            "(401132, 2)\n",
            "Epoch 1/20\n",
            "12536/12536 [==============================] - 42s 3ms/step - loss: 0.6704 - accuracy: 0.6245 - val_loss: 0.6555 - val_accuracy: 0.6357\n",
            "Epoch 2/20\n",
            "12536/12536 [==============================] - 40s 3ms/step - loss: 0.6563 - accuracy: 0.6330 - val_loss: 0.6553 - val_accuracy: 0.6357\n",
            "Epoch 3/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6559 - accuracy: 0.6340 - val_loss: 0.6553 - val_accuracy: 0.6357\n",
            "Epoch 4/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6573 - accuracy: 0.6314 - val_loss: 0.6554 - val_accuracy: 0.6356\n",
            "Epoch 5/20\n",
            "12536/12536 [==============================] - 40s 3ms/step - loss: 0.6564 - accuracy: 0.6332 - val_loss: 0.6556 - val_accuracy: 0.6356\n",
            "Epoch 6/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6557 - accuracy: 0.6343 - val_loss: 0.6555 - val_accuracy: 0.6357\n",
            "Epoch 7/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6568 - accuracy: 0.6324 - val_loss: 0.6550 - val_accuracy: 0.6357\n",
            "Epoch 8/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6560 - accuracy: 0.6332 - val_loss: 0.6550 - val_accuracy: 0.6357\n",
            "Epoch 9/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6559 - accuracy: 0.6339 - val_loss: 0.6550 - val_accuracy: 0.6357\n",
            "Epoch 10/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6560 - accuracy: 0.6335 - val_loss: 0.6553 - val_accuracy: 0.6357\n",
            "Epoch 11/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6571 - accuracy: 0.6320 - val_loss: 0.6554 - val_accuracy: 0.6357\n",
            "Epoch 12/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6559 - accuracy: 0.6336 - val_loss: 0.6551 - val_accuracy: 0.6357\n",
            "Epoch 13/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6567 - accuracy: 0.6325 - val_loss: 0.6551 - val_accuracy: 0.6357\n",
            "Epoch 14/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6560 - accuracy: 0.6331 - val_loss: 0.6551 - val_accuracy: 0.6357\n",
            "Epoch 15/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6567 - accuracy: 0.6321 - val_loss: 0.6549 - val_accuracy: 0.6357\n",
            "Epoch 16/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6564 - accuracy: 0.6326 - val_loss: 0.6551 - val_accuracy: 0.6357\n",
            "Epoch 17/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6562 - accuracy: 0.6333 - val_loss: 0.6549 - val_accuracy: 0.6357\n",
            "Epoch 18/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6557 - accuracy: 0.6338 - val_loss: 0.6549 - val_accuracy: 0.6357\n",
            "Epoch 19/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6560 - accuracy: 0.6329 - val_loss: 0.6549 - val_accuracy: 0.6357\n",
            "Epoch 20/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6554 - accuracy: 0.6343 - val_loss: 0.6550 - val_accuracy: 0.6357\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UvqmP52ENe2R",
        "outputId": "7ce19c9d-cff9-43aa-cac2-42fa3eafcdab"
      },
      "source": [
        "# Simple MLP for classification with Dimensionality resduction and Standardizing input\r\n",
        "import pandas as pd\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense\r\n",
        "from sklearn.preprocessing import StandardScaler\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "y = y.reshape(-1,1)\r\n",
        "\r\n",
        "# Scaling the data with StandardScaler\r\n",
        "\r\n",
        "scaler =StandardScaler()\r\n",
        "scaledX = scaler.fit_transform(X)\r\n",
        "print(scaledX)\r\n",
        "\r\n",
        "# one hot encoding the y\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "\r\n",
        "#Split into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(scaledX,y,test_size = 0.25)\r\n",
        "print(X_train.shape)\r\n",
        "print(y_train.shape)\r\n",
        "\r\n",
        "# Neural network\r\n",
        "model = Sequential()\r\n",
        "model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.25))\r\n",
        "model.add(Dense(2, activation='sigmoid'))\r\n",
        "# compile the keras model\r\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=20, batch_size=32, validation_data=(X_test, y_test),  verbose=1)\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[-0.25985153 -0.29290773  0.01628602 ... -0.53380379  0.47878796\n",
            "  -0.08700201]\n",
            " [-0.08538284 -0.29290773  0.01628602 ... -0.63418425 -0.51996429\n",
            "  -0.08700201]\n",
            " [-0.03406852 -0.29290773  0.01628602 ... -0.27281461  0.17147957\n",
            "  -0.08700201]\n",
            " ...\n",
            " [-0.24445723 -0.29290773  0.01628602 ... -0.14232001 -2.13333329\n",
            "  -0.08700201]\n",
            " [ 0.58170333 -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]\n",
            " [-0.2547201  -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]]\n",
            "(401132, 14)\n",
            "(401132, 2)\n",
            "Epoch 1/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6750 - accuracy: 0.6220 - val_loss: 0.6545 - val_accuracy: 0.6363\n",
            "Epoch 2/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6571 - accuracy: 0.6322 - val_loss: 0.6547 - val_accuracy: 0.6363\n",
            "Epoch 3/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6559 - accuracy: 0.6336 - val_loss: 0.6548 - val_accuracy: 0.6363\n",
            "Epoch 4/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6565 - accuracy: 0.6330 - val_loss: 0.6545 - val_accuracy: 0.6362\n",
            "Epoch 5/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6562 - accuracy: 0.6333 - val_loss: 0.6546 - val_accuracy: 0.6364\n",
            "Epoch 6/20\n",
            "12536/12536 [==============================] - 40s 3ms/step - loss: 0.6558 - accuracy: 0.6339 - val_loss: 0.6547 - val_accuracy: 0.6363\n",
            "Epoch 7/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6567 - accuracy: 0.6326 - val_loss: 0.6547 - val_accuracy: 0.6362\n",
            "Epoch 8/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6557 - accuracy: 0.6343 - val_loss: 0.6545 - val_accuracy: 0.6363\n",
            "Epoch 9/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6563 - accuracy: 0.6323 - val_loss: 0.6546 - val_accuracy: 0.6360\n",
            "Epoch 10/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6569 - accuracy: 0.6312 - val_loss: 0.6545 - val_accuracy: 0.6363\n",
            "Epoch 11/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6566 - accuracy: 0.6323 - val_loss: 0.6543 - val_accuracy: 0.6363\n",
            "Epoch 12/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6556 - accuracy: 0.6340 - val_loss: 0.6544 - val_accuracy: 0.6362\n",
            "Epoch 13/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6561 - accuracy: 0.6331 - val_loss: 0.6546 - val_accuracy: 0.6363\n",
            "Epoch 14/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6560 - accuracy: 0.6333 - val_loss: 0.6544 - val_accuracy: 0.6362\n",
            "Epoch 15/20\n",
            "12536/12536 [==============================] - 37s 3ms/step - loss: 0.6552 - accuracy: 0.6350 - val_loss: 0.6543 - val_accuracy: 0.6363\n",
            "Epoch 16/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6562 - accuracy: 0.6326 - val_loss: 0.6546 - val_accuracy: 0.6363\n",
            "Epoch 17/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6564 - accuracy: 0.6330 - val_loss: 0.6543 - val_accuracy: 0.6364\n",
            "Epoch 18/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6557 - accuracy: 0.6334 - val_loss: 0.6544 - val_accuracy: 0.6363\n",
            "Epoch 19/20\n",
            "12536/12536 [==============================] - 39s 3ms/step - loss: 0.6559 - accuracy: 0.6338 - val_loss: 0.6544 - val_accuracy: 0.6364\n",
            "Epoch 20/20\n",
            "12536/12536 [==============================] - 38s 3ms/step - loss: 0.6556 - accuracy: 0.6340 - val_loss: 0.6543 - val_accuracy: 0.6364\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UkMO3P6IN6PO"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t9VP1bGo7Q7E",
        "outputId": "94c37f77-db99-48b1-a7f7-cb41583f9659"
      },
      "source": [
        "# MLP for training dataset with grid search via sklearn\r\n",
        "import pandas as pd\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense, BatchNormalization, Dropout\r\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\r\n",
        "from sklearn.model_selection import GridSearchCV\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from sklearn.preprocessing import StandardScaler\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "import numpy\r\n",
        "\r\n",
        "# Function to create model, required for KerasClassifier\r\n",
        "def create_model(optimizer='rmsprop'):\r\n",
        "\t# create model\r\n",
        "  model = Sequential()\r\n",
        "  model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "  model.add(BatchNormalization())\r\n",
        "  model.add(Dropout(0.25))\r\n",
        "  model.add(Dense(2, activation='sigmoid'))\r\n",
        "\t# Compile model\r\n",
        "  model.compile(loss='binary_crossentropy', optimizer=optimizer, metrics=['accuracy'])\r\n",
        "  return model\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = ppd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "# Taking a small dataset\r\n",
        "dataset = dataset.sample(frac = .05)\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "y = y.reshape(-1,1)\r\n",
        "\r\n",
        "# Scaling the data with StandardScaler\r\n",
        "\r\n",
        "scaler =StandardScaler()\r\n",
        "scaledX = scaler.fit_transform(X)\r\n",
        "print(scaledX)\r\n",
        "\r\n",
        "# one hot encoding the y\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "\r\n",
        "#Split into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(scaledX,y,test_size = 0.25)\r\n",
        "print(X_test.shape)\r\n",
        "print(y_test.shape)\r\n",
        "\r\n",
        "# create model\r\n",
        "model = KerasClassifier(build_fn=create_model, verbose=0)\r\n",
        "# grid search epochs, batch size and optimizer\r\n",
        "optimizers = ['rmsprop', 'adam']\r\n",
        "epochs = [30, 40]\r\n",
        "batches = [32,64]\r\n",
        "param_grid = dict(optimizer=optimizers, epochs=epochs, batch_size=batches)\r\n",
        "grid = GridSearchCV(estimator=model, param_grid=param_grid)\r\n",
        "grid_result = grid.fit(X_train, y_train, validation_data=(X_test, y_test),  verbose=0)\r\n",
        "\r\n",
        "# summarize results\r\n",
        "print(\"Best: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))\r\n",
        "means = grid_result.cv_results_['mean_test_score']\r\n",
        "stds = grid_result.cv_results_['std_test_score']\r\n",
        "params = grid_result.cv_results_['params']\r\n",
        "for mean, stdev, param in zip(means, stds, params):\r\n",
        "\tprint(\"%f (%f) with: %r\" % (mean, stdev, param))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[-0.24796501 -0.29465989  0.013799   ... -0.53832527  0.63147466\n",
            "  -0.08300795]\n",
            " [ 0.03677989 -0.29465989  0.013799   ... -0.24140889  1.6317783\n",
            "  -0.08300795]\n",
            " [-0.23323683 -0.29465989  0.013799   ...  0.25003891 -0.13798968\n",
            "  -0.08300795]\n",
            " ...\n",
            " [-0.24060092 -0.29465989  0.013799   ... -0.00592349  1.55483186\n",
            "  -0.08300795]\n",
            " [-0.18414288 -0.29465989  0.013799   ... -0.2106934  -0.06104325\n",
            "  -0.08300795]\n",
            " [-0.23323683 -0.29465989  0.013799   ... -0.04687747  0.01590318\n",
            "  -0.08300795]]\n",
            "(6686, 14)\n",
            "(6686, 2)\n",
            "Best: 0.640756 using {'batch_size': 64, 'epochs': 40, 'optimizer': 'adam'}\n",
            "0.640706 (0.004110) with: {'batch_size': 32, 'epochs': 30, 'optimizer': 'rmsprop'}\n",
            "0.640207 (0.004012) with: {'batch_size': 32, 'epochs': 30, 'optimizer': 'adam'}\n",
            "0.640307 (0.003567) with: {'batch_size': 32, 'epochs': 40, 'optimizer': 'rmsprop'}\n",
            "0.640407 (0.004308) with: {'batch_size': 32, 'epochs': 40, 'optimizer': 'adam'}\n",
            "0.640357 (0.003818) with: {'batch_size': 64, 'epochs': 30, 'optimizer': 'rmsprop'}\n",
            "0.639659 (0.004590) with: {'batch_size': 64, 'epochs': 30, 'optimizer': 'adam'}\n",
            "0.640506 (0.003805) with: {'batch_size': 64, 'epochs': 40, 'optimizer': 'rmsprop'}\n",
            "0.640756 (0.004276) with: {'batch_size': 64, 'epochs': 40, 'optimizer': 'adam'}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gbzvSLXQySIC"
      },
      "source": [
        "2. Convolutional Nerural Network CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xF_jUiAL7Qwi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ed4a0406-365d-4aa4-87f2-b8a272640d70"
      },
      "source": [
        "# Training CNN with Standard Scaler with dimesnsionality reduction\r\n",
        "\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense, Conv1D\r\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from sklearn.preprocessing  import StandardScaler\r\n",
        "from keras.layers import BatchNormalization, Dropout, Flatten\r\n",
        "from keras.optimizers import Adam\r\n",
        "import numpy\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "y = y.reshape(-1,1)\r\n",
        "\r\n",
        "# Scaling the data with StandardScaler\r\n",
        "\r\n",
        "scaler =StandardScaler()\r\n",
        "scaledX = scaler.fit_transform(X)\r\n",
        "print(scaledX)\r\n",
        "\r\n",
        "# one hot encoding the y\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "\r\n",
        "\r\n",
        "X_train, X_test, y_train, y_test = train_test_split(scaledX, y, test_size = 0.2, random_state = 0)\r\n",
        "print(X_train.shape, X_test.shape)\r\n",
        "\r\n",
        "X_train = X_train.reshape(X_train.shape[0], X_train.shape[1], 1)\r\n",
        "X_test = X_test.reshape(X_test.shape[0], X_test.shape[1], 1)\r\n",
        "\r\n",
        "X_train.shape, X_test.shape\r\n",
        "\r\n",
        "\r\n",
        "epochs = 20\r\n",
        "model = Sequential()\r\n",
        "model.add(Conv1D(32, 2, activation='relu', input_shape = X_train[0].shape))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.2))\r\n",
        "\r\n",
        "model.add(Conv1D(64, 2, activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.5))\r\n",
        "\r\n",
        "model.add(Flatten())\r\n",
        "model.add(Dense(64, activation='relu'))\r\n",
        "model.add(Dropout(0.5))\r\n",
        "\r\n",
        "model.add(Dense(2, activation='sigmoid'))\r\n",
        "\r\n",
        "model.summary()\r\n",
        "model.compile(optimizer=Adam(lr=0.0001), loss = 'binary_crossentropy', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=epochs, validation_data=(X_test, y_test), verbose=1)\r\n",
        "\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[-0.25985153 -0.29290773  0.01628602 ... -0.53380379  0.47878796\n",
            "  -0.08700201]\n",
            " [-0.08538284 -0.29290773  0.01628602 ... -0.63418425 -0.51996429\n",
            "  -0.08700201]\n",
            " [-0.03406852 -0.29290773  0.01628602 ... -0.27281461  0.17147957\n",
            "  -0.08700201]\n",
            " ...\n",
            " [-0.24445723 -0.29290773  0.01628602 ... -0.14232001 -2.13333329\n",
            "  -0.08700201]\n",
            " [ 0.58170333 -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]\n",
            " [-0.2547201  -0.29290773  0.01628602 ... -0.4033092  -1.21140814\n",
            "  -0.08700201]]\n",
            "(427874, 14) (106969, 14)\n",
            "Model: \"sequential_113\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_1 (Conv1D)            (None, 13, 32)            96        \n",
            "_________________________________________________________________\n",
            "batch_normalization_112 (Bat (None, 13, 32)            128       \n",
            "_________________________________________________________________\n",
            "dropout_112 (Dropout)        (None, 13, 32)            0         \n",
            "_________________________________________________________________\n",
            "conv1d_2 (Conv1D)            (None, 12, 64)            4160      \n",
            "_________________________________________________________________\n",
            "batch_normalization_113 (Bat (None, 12, 64)            256       \n",
            "_________________________________________________________________\n",
            "dropout_113 (Dropout)        (None, 12, 64)            0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 768)               0         \n",
            "_________________________________________________________________\n",
            "dense_224 (Dense)            (None, 64)                49216     \n",
            "_________________________________________________________________\n",
            "dropout_114 (Dropout)        (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_225 (Dense)            (None, 2)                 130       \n",
            "=================================================================\n",
            "Total params: 53,986\n",
            "Trainable params: 53,794\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "13372/13372 [==============================] - 57s 4ms/step - loss: 0.7571 - accuracy: 0.6031 - val_loss: 0.6575 - val_accuracy: 0.6264\n",
            "Epoch 2/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6591 - accuracy: 0.6303 - val_loss: 0.6559 - val_accuracy: 0.6332\n",
            "Epoch 3/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6578 - accuracy: 0.6304 - val_loss: 0.6558 - val_accuracy: 0.6338\n",
            "Epoch 4/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6567 - accuracy: 0.6324 - val_loss: 0.6556 - val_accuracy: 0.6345\n",
            "Epoch 5/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6559 - accuracy: 0.6332 - val_loss: 0.6563 - val_accuracy: 0.6309\n",
            "Epoch 6/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6560 - accuracy: 0.6328 - val_loss: 0.6556 - val_accuracy: 0.6344\n",
            "Epoch 7/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6559 - accuracy: 0.6329 - val_loss: 0.6559 - val_accuracy: 0.6341\n",
            "Epoch 8/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6570 - accuracy: 0.6308 - val_loss: 0.6554 - val_accuracy: 0.6346\n",
            "Epoch 9/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6553 - accuracy: 0.6336 - val_loss: 0.6554 - val_accuracy: 0.6344\n",
            "Epoch 10/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6561 - accuracy: 0.6331 - val_loss: 0.6556 - val_accuracy: 0.6341\n",
            "Epoch 11/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6556 - accuracy: 0.6335 - val_loss: 0.6552 - val_accuracy: 0.6347\n",
            "Epoch 12/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6554 - accuracy: 0.6335 - val_loss: 0.6552 - val_accuracy: 0.6347\n",
            "Epoch 13/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6555 - accuracy: 0.6330 - val_loss: 0.6554 - val_accuracy: 0.6344\n",
            "Epoch 14/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6555 - accuracy: 0.6328 - val_loss: 0.6556 - val_accuracy: 0.6340\n",
            "Epoch 15/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6561 - accuracy: 0.6322 - val_loss: 0.6552 - val_accuracy: 0.6347\n",
            "Epoch 16/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6558 - accuracy: 0.6325 - val_loss: 0.6552 - val_accuracy: 0.6348\n",
            "Epoch 17/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6564 - accuracy: 0.6312 - val_loss: 0.6551 - val_accuracy: 0.6348\n",
            "Epoch 18/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6559 - accuracy: 0.6325 - val_loss: 0.6552 - val_accuracy: 0.6347\n",
            "Epoch 19/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6559 - accuracy: 0.6325 - val_loss: 0.6552 - val_accuracy: 0.6346\n",
            "Epoch 20/20\n",
            "13372/13372 [==============================] - 57s 4ms/step - loss: 0.6553 - accuracy: 0.6337 - val_loss: 0.6552 - val_accuracy: 0.6348\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4cMcSnFKcsqP"
      },
      "source": [
        "def plot_learningCurve(history, epoch):\r\n",
        "  # Plot training & validation accuracy values\r\n",
        "  epoch_range = range(1, epoch+1)\r\n",
        "  plt.plot(epoch_range, history.history['accuracy'])\r\n",
        "  plt.plot(epoch_range, history.history['val_accuracy'])\r\n",
        "  plt.title('Model accuracy')\r\n",
        "  plt.ylabel('Accuracy')\r\n",
        "  plt.xlabel('Epoch')\r\n",
        "  plt.legend(['Train', 'Val'], loc='upper left')\r\n",
        "  plt.show()\r\n",
        "\r\n",
        "  # Plot training & validation loss values\r\n",
        "  plt.plot(epoch_range, history.history['loss'])\r\n",
        "  plt.plot(epoch_range, history.history['val_loss'])\r\n",
        "  plt.title('Model loss')\r\n",
        "  plt.ylabel('Loss')\r\n",
        "  plt.xlabel('Epoch')\r\n",
        "  plt.legend(['Train', 'Val'], loc='upper left')\r\n",
        "  plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "clM6M3p1c2PU",
        "outputId": "2d580e87-ca11-40d4-999e-d3d17660f0d8"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "plot_learningCurve(history, epochs)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bPE-V1Nv7Qaf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4056ed49-83a3-434c-b2d3-874c628cac8c"
      },
      "source": [
        "# Training CNN with MinMax Scaler with dimesnsionality reduction\r\n",
        "\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense, Conv1D\r\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from sklearn.preprocessing  import MinMaxScaler\r\n",
        "from keras.layers import BatchNormalization, Dropout, Flatten\r\n",
        "from keras.optimizers import Adam\r\n",
        "import numpy\r\n",
        "\r\n",
        "\r\n",
        "# load the dataset\r\n",
        "dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "\r\n",
        "# Droping the Duplicate row \r\n",
        "dataset = dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "# Droping some of the column from input \"Dimensionality Reduction\"\r\n",
        "X_input = dataset.drop(['Provider','PotentialFraud','NoOfMonths_PartBCov','RenalDiseaseIndicator','ChronicCond_KidneyDisease',\r\n",
        "                    'ChronicCond_Diabetes','ChronicCond_IschemicHeart','ChronicCond_Heartfailure','ChronicCond_rheumatoidarthritis',\r\n",
        "                    'ChronicCond_Cancer','Gender','ChronicCond_ObstrPulmonary','ChronicCond_Alzheimer'], axis = 1)\r\n",
        "y_output = dataset.PotentialFraud\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "X = X_input.values\r\n",
        "y = y_output.values\r\n",
        "y = y.reshape(-1,1)\r\n",
        "\r\n",
        "# Scaling the data with MinMaxScaler\r\n",
        "\r\n",
        "scaler =MinMaxScaler()\r\n",
        "scaledX = scaler.fit_transform(X)\r\n",
        "print(scaledX)\r\n",
        "\r\n",
        "# one hot encoding the y\r\n",
        "ohe = OneHotEncoder()\r\n",
        "y = ohe.fit_transform(y).toarray()\r\n",
        "\r\n",
        "\r\n",
        "X_train, X_test, y_train, y_test = train_test_split(scaledX, y, test_size = 0.2, random_state = 0)\r\n",
        "print(X_train.shape, X_test.shape)\r\n",
        "\r\n",
        "X_train = X_train.reshape(X_train.shape[0], X_train.shape[1], 1)\r\n",
        "X_test = X_test.reshape(X_test.shape[0], X_test.shape[1], 1)\r\n",
        "\r\n",
        "X_train.shape, X_test.shape\r\n",
        "\r\n",
        "\r\n",
        "epochs = 20\r\n",
        "model = Sequential()\r\n",
        "model.add(Conv1D(32, 2, activation='relu', input_shape = X_train[0].shape))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.2))\r\n",
        "\r\n",
        "model.add(Conv1D(64, 2, activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.5))\r\n",
        "\r\n",
        "model.add(Flatten())\r\n",
        "model.add(Dense(64, activation='relu'))\r\n",
        "model.add(Dropout(0.5))\r\n",
        "\r\n",
        "model.add(Dense(2, activation='sigmoid'))\r\n",
        "\r\n",
        "model.summary()\r\n",
        "model.compile(optimizer=Adam(lr=0.0001), loss = 'binary_crossentropy', metrics=['accuracy'])\r\n",
        "history = model.fit(X_train, y_train, epochs=epochs, validation_data=(X_test, y_test), verbose=1)\r\n",
        "\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1.60000000e-04 0.00000000e+00 1.71428571e-01 ... 7.94797688e-03\n",
            "  7.20000000e-01 0.00000000e+00]\n",
            " [5.60000000e-03 0.00000000e+00 1.71428571e-01 ... 7.22543353e-04\n",
            "  5.46666667e-01 0.00000000e+00]\n",
            " [7.20000000e-03 0.00000000e+00 1.71428571e-01 ... 2.67341040e-02\n",
            "  6.66666667e-01 0.00000000e+00]\n",
            " ...\n",
            " [6.40000000e-04 0.00000000e+00 1.71428571e-01 ... 3.61271676e-02\n",
            "  2.66666667e-01 0.00000000e+00]\n",
            " [2.64000000e-02 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]\n",
            " [3.20000000e-04 0.00000000e+00 1.71428571e-01 ... 1.73410405e-02\n",
            "  4.26666667e-01 0.00000000e+00]]\n",
            "(427874, 14) (106969, 14)\n",
            "Model: \"sequential_114\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_3 (Conv1D)            (None, 13, 32)            96        \n",
            "_________________________________________________________________\n",
            "batch_normalization_114 (Bat (None, 13, 32)            128       \n",
            "_________________________________________________________________\n",
            "dropout_115 (Dropout)        (None, 13, 32)            0         \n",
            "_________________________________________________________________\n",
            "conv1d_4 (Conv1D)            (None, 12, 64)            4160      \n",
            "_________________________________________________________________\n",
            "batch_normalization_115 (Bat (None, 12, 64)            256       \n",
            "_________________________________________________________________\n",
            "dropout_116 (Dropout)        (None, 12, 64)            0         \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 768)               0         \n",
            "_________________________________________________________________\n",
            "dense_226 (Dense)            (None, 64)                49216     \n",
            "_________________________________________________________________\n",
            "dropout_117 (Dropout)        (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_227 (Dense)            (None, 2)                 130       \n",
            "=================================================================\n",
            "Total params: 53,986\n",
            "Trainable params: 53,794\n",
            "Non-trainable params: 192\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "13372/13372 [==============================] - 57s 4ms/step - loss: 0.6862 - accuracy: 0.6188 - val_loss: 0.6571 - val_accuracy: 0.6345\n",
            "Epoch 2/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6588 - accuracy: 0.6303 - val_loss: 0.6567 - val_accuracy: 0.6345\n",
            "Epoch 3/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6572 - accuracy: 0.6317 - val_loss: 0.6560 - val_accuracy: 0.6345\n",
            "Epoch 4/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6572 - accuracy: 0.6308 - val_loss: 0.6560 - val_accuracy: 0.6345\n",
            "Epoch 5/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6566 - accuracy: 0.6315 - val_loss: 0.6557 - val_accuracy: 0.6345\n",
            "Epoch 6/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6572 - accuracy: 0.6307 - val_loss: 0.6559 - val_accuracy: 0.6345\n",
            "Epoch 7/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6566 - accuracy: 0.6316 - val_loss: 0.6556 - val_accuracy: 0.6345\n",
            "Epoch 8/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6570 - accuracy: 0.6306 - val_loss: 0.6556 - val_accuracy: 0.6345\n",
            "Epoch 9/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6569 - accuracy: 0.6309 - val_loss: 0.6556 - val_accuracy: 0.6345\n",
            "Epoch 10/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6560 - accuracy: 0.6326 - val_loss: 0.6555 - val_accuracy: 0.6345\n",
            "Epoch 11/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6564 - accuracy: 0.6322 - val_loss: 0.6555 - val_accuracy: 0.6345\n",
            "Epoch 12/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6565 - accuracy: 0.6315 - val_loss: 0.6556 - val_accuracy: 0.6345\n",
            "Epoch 13/20\n",
            "13372/13372 [==============================] - 55s 4ms/step - loss: 0.6563 - accuracy: 0.6324 - val_loss: 0.6555 - val_accuracy: 0.6345\n",
            "Epoch 14/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6561 - accuracy: 0.6323 - val_loss: 0.6556 - val_accuracy: 0.6346\n",
            "Epoch 15/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6561 - accuracy: 0.6323 - val_loss: 0.6555 - val_accuracy: 0.6345\n",
            "Epoch 16/20\n",
            "13372/13372 [==============================] - 57s 4ms/step - loss: 0.6562 - accuracy: 0.6322 - val_loss: 0.6554 - val_accuracy: 0.6346\n",
            "Epoch 17/20\n",
            "13372/13372 [==============================] - 57s 4ms/step - loss: 0.6559 - accuracy: 0.6324 - val_loss: 0.6555 - val_accuracy: 0.6345\n",
            "Epoch 18/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6563 - accuracy: 0.6320 - val_loss: 0.6560 - val_accuracy: 0.6334\n",
            "Epoch 19/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6566 - accuracy: 0.6311 - val_loss: 0.6554 - val_accuracy: 0.6345\n",
            "Epoch 20/20\n",
            "13372/13372 [==============================] - 56s 4ms/step - loss: 0.6555 - accuracy: 0.6331 - val_loss: 0.6556 - val_accuracy: 0.6344\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D74s2uGv7QXg"
      },
      "source": [
        "def plot_learningCurve(history, epoch):\r\n",
        "  # Plot training & validation accuracy values\r\n",
        "  epoch_range = range(1, epoch+1)\r\n",
        "  plt.plot(epoch_range, history.history['accuracy'])\r\n",
        "  plt.plot(epoch_range, history.history['val_accuracy'])\r\n",
        "  plt.title('Model accuracy')\r\n",
        "  plt.ylabel('Accuracy')\r\n",
        "  plt.xlabel('Epoch')\r\n",
        "  plt.legend(['Train', 'Val'], loc='upper left')\r\n",
        "  plt.show()\r\n",
        "\r\n",
        "  # Plot training & validation loss values\r\n",
        "  plt.plot(epoch_range, history.history['loss'])\r\n",
        "  plt.plot(epoch_range, history.history['val_loss'])\r\n",
        "  plt.title('Model loss')\r\n",
        "  plt.ylabel('Loss')\r\n",
        "  plt.xlabel('Epoch')\r\n",
        "  plt.legend(['Train', 'Val'], loc='upper left')\r\n",
        "  plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "qj4JHatTqXwI",
        "outputId": "7fad4e7d-aa35-4c40-fb31-225d5128011d"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "plot_learningCurve(history, epochs)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z6OFjI4rQ6JV"
      },
      "source": [
        "## Keras Classifier with model saving and Prediction"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EgW23t8P_yqP"
      },
      "source": [
        "#### Prepcocessing and model training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jyPMk-QyPCwi",
        "outputId": "db41b091-23a6-492b-a2b5-151c58f86c67"
      },
      "source": [
        "# Simple MLP for classification with Prediction and Model\r\n",
        "import pandas as pd\r\n",
        "from sklearn.model_selection import train_test_split\r\n",
        "from keras.models import Sequential\r\n",
        "from keras.layers import Dense, BatchNormalization, Dropout\r\n",
        "from keras.callbacks import ModelCheckpoint, EarlyStopping\r\n",
        "from sklearn.preprocessing import MinMaxScaler\r\n",
        "from sklearn.preprocessing import OneHotEncoder\r\n",
        "\r\n",
        "# load the dataset for training \r\n",
        "train_dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/train_data.csv')\r\n",
        "\r\n",
        "# Droping the Duplicate row \r\n",
        "train_dataset = train_dataset.drop_duplicates(subset = None,inplace= False )\r\n",
        "\r\n",
        "#Changing pandas dataframe into input and output numpy array\r\n",
        "input = train_dataset.iloc[:,2:28].values.astype('float32')\r\n",
        "output = train_dataset.iloc[:,1:2].values.astype('float32')\r\n",
        "\r\n",
        "# Scaled the input with MinMaxScaler\r\n",
        "input_scaler = MinMaxScaler(feature_range=(0, 1))\r\n",
        "scaled_input = input_scaler.fit_transform(input)\r\n",
        "\r\n",
        "# one hot encoding the output of training dataset\r\n",
        "output_encoder = OneHotEncoder()\r\n",
        "encoded_output = output_encoder.fit_transform(output).toarray()\r\n",
        "\r\n",
        "print(scaled_input.shape)\r\n",
        "print(encoded_output.shape)\r\n",
        "\r\n",
        "input_dim = scaled_input.shape[-1]\r\n",
        "print(input_dim)\r\n",
        "\r\n",
        "#Split into train test set\r\n",
        "X_train,X_test,y_train,y_test = train_test_split(scaled_input, encoded_output, test_size = 0.25)\r\n",
        "\r\n",
        "\r\n",
        "# Neural network\r\n",
        "model = Sequential()\r\n",
        "model.add(Dense(36, input_dim=X_train.shape[-1], activation='relu'))\r\n",
        "model.add(Dense(18, activation='relu'))\r\n",
        "model.add(BatchNormalization())\r\n",
        "model.add(Dropout(0.25))\r\n",
        "model.add(Dense(2, activation='softmax'))\r\n",
        "\r\n",
        "# compile the keras model\r\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\r\n",
        "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=20)\r\n",
        "mc = ModelCheckpoint('/content/drive/My Drive/ADM Datasets/mlp_model.h5', monitor='val_loss', mode='min', verbose=1, save_best_only=True)\r\n",
        "history = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=20, batch_size=32, verbose=2, callbacks = [es,mc])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(534843, 25)\n",
            "(534843, 2)\n",
            "25\n",
            "Epoch 1/20\n",
            "12536/12536 - 35s - loss: 0.6598 - accuracy: 0.6299 - val_loss: 0.6554 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00001: val_loss improved from inf to 0.65544, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 2/20\n",
            "12536/12536 - 33s - loss: 0.6565 - accuracy: 0.6332 - val_loss: 0.6552 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00002: val_loss improved from 0.65544 to 0.65518, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 3/20\n",
            "12536/12536 - 33s - loss: 0.6565 - accuracy: 0.6333 - val_loss: 0.6550 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00003: val_loss improved from 0.65518 to 0.65502, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 4/20\n",
            "12536/12536 - 33s - loss: 0.6563 - accuracy: 0.6333 - val_loss: 0.6553 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00004: val_loss did not improve from 0.65502\n",
            "Epoch 5/20\n",
            "12536/12536 - 33s - loss: 0.6563 - accuracy: 0.6333 - val_loss: 0.6550 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00005: val_loss improved from 0.65502 to 0.65499, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 6/20\n",
            "12536/12536 - 33s - loss: 0.6562 - accuracy: 0.6334 - val_loss: 0.6548 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00006: val_loss improved from 0.65499 to 0.65485, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 7/20\n",
            "12536/12536 - 35s - loss: 0.6561 - accuracy: 0.6332 - val_loss: 0.6549 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00007: val_loss did not improve from 0.65485\n",
            "Epoch 8/20\n",
            "12536/12536 - 33s - loss: 0.6558 - accuracy: 0.6333 - val_loss: 0.6549 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00008: val_loss did not improve from 0.65485\n",
            "Epoch 9/20\n",
            "12536/12536 - 33s - loss: 0.6559 - accuracy: 0.6332 - val_loss: 0.6552 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00009: val_loss did not improve from 0.65485\n",
            "Epoch 10/20\n",
            "12536/12536 - 35s - loss: 0.6560 - accuracy: 0.6331 - val_loss: 0.6548 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00010: val_loss improved from 0.65485 to 0.65483, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 11/20\n",
            "12536/12536 - 34s - loss: 0.6559 - accuracy: 0.6331 - val_loss: 0.6549 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00011: val_loss did not improve from 0.65483\n",
            "Epoch 12/20\n",
            "12536/12536 - 34s - loss: 0.6558 - accuracy: 0.6335 - val_loss: 0.6550 - val_accuracy: 0.6356\n",
            "\n",
            "Epoch 00012: val_loss did not improve from 0.65483\n",
            "Epoch 13/20\n",
            "12536/12536 - 34s - loss: 0.6557 - accuracy: 0.6334 - val_loss: 0.6552 - val_accuracy: 0.6356\n",
            "\n",
            "Epoch 00013: val_loss did not improve from 0.65483\n",
            "Epoch 14/20\n",
            "12536/12536 - 33s - loss: 0.6558 - accuracy: 0.6334 - val_loss: 0.6547 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00014: val_loss improved from 0.65483 to 0.65465, saving model to /content/drive/My Drive/ADM Datasets/mlp_model.h5\n",
            "Epoch 15/20\n",
            "12536/12536 - 33s - loss: 0.6557 - accuracy: 0.6333 - val_loss: 0.6549 - val_accuracy: 0.6356\n",
            "\n",
            "Epoch 00015: val_loss did not improve from 0.65465\n",
            "Epoch 16/20\n",
            "12536/12536 - 34s - loss: 0.6557 - accuracy: 0.6332 - val_loss: 0.6547 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00016: val_loss did not improve from 0.65465\n",
            "Epoch 17/20\n",
            "12536/12536 - 33s - loss: 0.6555 - accuracy: 0.6333 - val_loss: 0.6547 - val_accuracy: 0.6356\n",
            "\n",
            "Epoch 00017: val_loss did not improve from 0.65465\n",
            "Epoch 18/20\n",
            "12536/12536 - 34s - loss: 0.6556 - accuracy: 0.6334 - val_loss: 0.6547 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00018: val_loss did not improve from 0.65465\n",
            "Epoch 19/20\n",
            "12536/12536 - 35s - loss: 0.6555 - accuracy: 0.6335 - val_loss: 0.6548 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00019: val_loss did not improve from 0.65465\n",
            "Epoch 20/20\n",
            "12536/12536 - 34s - loss: 0.6554 - accuracy: 0.6336 - val_loss: 0.6547 - val_accuracy: 0.6355\n",
            "\n",
            "Epoch 00020: val_loss did not improve from 0.65465\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ifxe7omv16ud"
      },
      "source": [
        "#### Model Evaluation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I-L9cEXa150L",
        "outputId": "324bdb9a-fb1b-4e04-9546-84f7664f25cf"
      },
      "source": [
        "loss_train, acc_train = model.evaluate(X_train, y_train, verbose = 0)\r\n",
        "loss_val, acc_val = model.evaluate(X_test, y_test, verbose = 0)\r\n",
        "print(\"The train loss is: {0} and the train accuracy is: {1}\".format(loss_train, acc_train))\r\n",
        "print(\"The validation loss is: {0} and the validation accuracy is: {1}\".format(loss_val, acc_val))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The train loss is: 0.6547264456748962 and the train accuracy is: 0.6347711086273193\n",
            "The validation loss is: 0.654743492603302 and the validation accuracy is: 0.6354600787162781\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B4t8UHE5zJ36"
      },
      "source": [
        "#### Accuracy, Precision, Recall, F1 Score, roc auc Curve and confusion marix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7bizyf0FyxvV",
        "outputId": "60e45653-0db1-4867-fdcc-648aab3c9c1f"
      },
      "source": [
        "from sklearn.metrics import accuracy_score\r\n",
        "from sklearn.metrics import precision_score\r\n",
        "from sklearn.metrics import recall_score\r\n",
        "from sklearn.metrics import f1_score\r\n",
        "from sklearn.metrics import cohen_kappa_score\r\n",
        "from sklearn.metrics import roc_auc_score\r\n",
        "from sklearn.metrics import confusion_matrix\r\n",
        "\r\n",
        "# predict probabilities for test set\r\n",
        "yhat_probs = model.predict(X_test, verbose=0)\r\n",
        "# predict crisp classes for test set\r\n",
        "yhat_classes = model.predict_classes(X_test, verbose=0)\r\n",
        "\r\n",
        "print(yhat_probs) # 2-D\r\n",
        "print(yhat_classes) # 1-D\r\n",
        "\r\n",
        "# reduce to 1d array\r\n",
        "yhat_probs = yhat_probs[:, 0]\r\n",
        "y_test_1D = output_encoder.inverse_transform(y_test) # Need to be in 1D\r\n",
        "\r\n",
        "print(yhat_probs)\r\n",
        "print(yhat_classes)\r\n",
        "print()\r\n",
        "\r\n",
        "\r\n",
        "# accuracy: (tp + tn) / (p + n)\r\n",
        "accuracy = accuracy_score(y_test_1D, yhat_classes)\r\n",
        "print('Accuracy: %f' % accuracy)\r\n",
        "# precision tp / (tp + fp)\r\n",
        "precision = precision_score(y_test_1D, yhat_classes)\r\n",
        "print('Precision: %f' % precision)\r\n",
        "# recall: tp / (tp + fn)\r\n",
        "recall = recall_score(y_test_1D, yhat_classes)\r\n",
        "print('Recall: %f' % recall)\r\n",
        "# f1: 2 tp / (2 tp + fp + fn)\r\n",
        "f1 = f1_score(y_test_1D, yhat_classes)\r\n",
        "print('F1 score: %f' % f1)\r\n",
        " \r\n",
        "# kappa\r\n",
        "kappa = cohen_kappa_score(y_test_1D, yhat_classes)\r\n",
        "print('Cohens kappa: %f' % kappa)\r\n",
        "# ROC AUC\r\n",
        "auc = roc_auc_score(y_test_1D, yhat_probs)\r\n",
        "print('ROC AUC: %f' % auc)\r\n",
        "# confusion matrix\r\n",
        "matrix = confusion_matrix(y_test_1D, yhat_classes)\r\n",
        "print(matrix)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/engine/sequential.py:450: UserWarning: `model.predict_classes()` is deprecated and will be removed after 2021-01-01. Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n",
            "  warnings.warn('`model.predict_classes()` is deprecated and '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "[[0.7558467  0.24415329]\n",
            " [0.40021473 0.5997853 ]\n",
            " [0.6733066  0.3266934 ]\n",
            " ...\n",
            " [0.6424555  0.35754448]\n",
            " [0.42232147 0.5776785 ]\n",
            " [0.7502461  0.24975388]]\n",
            "[0 1 0 ... 0 1 0]\n",
            "[0.7558467  0.40021473 0.6733066  ... 0.6424555  0.42232147 0.7502461 ]\n",
            "[0 1 0 ... 0 1 0]\n",
            "\n",
            "Accuracy: 0.635460\n",
            "Precision: 0.576935\n",
            "Recall: 0.112882\n",
            "F1 score: 0.188820\n",
            "Cohens kappa: 0.075040\n",
            "ROC AUC: 0.450578\n",
            "[[79295  4160]\n",
            " [44583  5673]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2QnCrPHU75-3"
      },
      "source": [
        "#### Visualization of Loss and Accuracy "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "xBxaPVwx5-Yn",
        "outputId": "f587e01a-90a2-4cf0-c321-5995cd4572ae"
      },
      "source": [
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "# plot loss during training\r\n",
        "\r\n",
        "plt.plot(history.history['loss'], label='Training loss')\r\n",
        "plt.plot(history.history['val_loss'], label='validation loss')\r\n",
        "plt.title('Training and Validation loss')\r\n",
        "plt.xlabel('Epochs')\r\n",
        "plt.ylabel('Loss')\r\n",
        "plt.legend()\r\n",
        "plt.show()\r\n",
        "\r\n",
        "# plot Accuracy during training\r\n",
        "plt.plot(history.history['accuracy'], label='Training Accuracy')\r\n",
        "plt.plot(history.history['val_accuracy'], label='validation Accuracy')\r\n",
        "plt.title('Training and Validation Accuracy')\r\n",
        "plt.xlabel('Epochs')\r\n",
        "plt.ylabel('Accuracy')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6ADsBWVh8Ru-"
      },
      "source": [
        "## Making Prediction on Saved Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F2YkPT158cZJ"
      },
      "source": [
        "#### Load the Saved Classifier or Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b5WvcUqdcBI3"
      },
      "source": [
        "from keras.models import load_model\r\n",
        "#Load the saved Model and do prediction\r\n",
        "\r\n",
        "classifier = load_model('/content/drive/My Drive/ADM Datasets/mlp_model.h5')\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nRIAfgq4AKn9"
      },
      "source": [
        "#### Prediction on test.csv"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mTEtVUcudhLn",
        "outputId": "723db9cc-d1bc-4b5a-cb51-b0440747b76d"
      },
      "source": [
        "# Now make prediction on saved Model for test dataset\r\n",
        "# load the dataset for testing the model\r\n",
        "\r\n",
        "pred_dataset = pd.read_csv('/content/drive/My Drive/ADM Datasets/test_data.csv')\r\n",
        "\r\n",
        "#Changing pandas dataframe to numpy array\r\n",
        "pred_input = pred_dataset.iloc[:,2:28].values.astype('float32')\r\n",
        "pred_output = pred_dataset.iloc[:,1:2].values.astype('float32')\r\n",
        "\r\n",
        "# Scaled the input and output for prediction\r\n",
        "pred_input_scaled = input_scaler.fit_transform(pred_input)\r\n",
        "pred_output_endcoded = output_encoder.fit_transform(pred_output).toarray()\r\n",
        "\r\n",
        "# evaluate the loss and accuracy for the prediction dataset\r\n",
        "pred_loss, pred_acc = classifier.evaluate(pred_input_scaled, pred_output_endcoded, verbose =0)\r\n",
        "print(\"The Predicted dataset loss is: {0} and Accuracy is: {1}\".format(pred_loss, pred_acc))\r\n",
        "\r\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The Predicted dataset loss is: 0.657598614692688 and Accuracy is: 0.6299875974655151\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q1sGJykr-ll5",
        "outputId": "75711f54-b4df-4b67-b0bd-d238f204bc19"
      },
      "source": [
        "# Prediction on input data\r\n",
        "\r\n",
        "prediction = classifier.predict(pred_input_scaled)\r\n",
        "inv_predict = output_encoder.inverse_transform(prediction).astype('int64')\r\n",
        "inv_actual = output_encoder.inverse_transform(pred_output_endcoded).astype('int64')\r\n",
        "for j in range(0,len(pred_output)):\r\n",
        "        print(\"Pred={0} Act={1}\\n\".format(inv_predict[j], inv_actual[j]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43mStreaming output truncated to the last 5000 lines.\u001b[0m\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[1] Act=[1]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[1] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n",
            "Pred=[0] Act=[0]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}