Learn.jl

WORK IN PROGRESS

Overview

Machine learning package with implementations of common algorithms and techniques.

Installation

Pkg.clone("https://github.com/equinetic/Learn.jl")
using Learn

Current Capabilities

• Supervised Algorithms
  • Linear Regression
  • Logistic Regression
• Utilities
  • Quantile and numeric binning
  • Feature scaling

Reexports LossFunctions, PenaltyFunctions, and MLDataUtils. See their pages for detailed documentation.

Planned

• Algorithms: Reimplementation of LightML.jl models
• Model Evaluation
  • Comprehensive build out of classification / regression cost functions
  • Extensible ROC Curves
  • Learning Curves
• Utilities
  • Variable binner (discretization)
  • Feature scaler
  • Pipeline
  • Easy partitioning

To do

See the TODO.md for a list of short to mid-term tasks.

Examples

• Linear Regression
• Logistic Regression