kind_to_fc_parameters return an empty Dataframe

[bhavara23]

I have a JSON file with 130 feature names along with the values.

I want to generate the features for my data with the attributes in the JSON file.
I am using kind_to_fc_parameters but its returning an empty dataframe.

features.json

{
  "partial_autocorrelation":[{"lag":9},{"lag":5},{"lag":6},{"lag":2},{"lag":1},{"lag":3},{"lag":8},{"lag":0},{"lag":7},{"lag":4}],

  "change_quantiles":[{"f_agg":"mean","isabs":"True","qh":0.8,"ql":0.0},{"f_agg":"var","isabs":"True","qh":1.0,"ql":0.4},{"f_agg":"mean","isabs":"False","qh":0.6,"ql":0.4},{"f_agg":"mean","isabs":"False","qh":0.8,"ql":0.6},
                      {"f_agg":"mean","isabs":"True","qh":0.8,"ql":0.6},{"f_agg":"mean","isabs":"False","qh":0.8,"ql":0.2},{"f_agg":"var","isabs":"False","qh":0.8,"ql":0.6},{"f_agg":"var","isabs":"False","qh":0.6,"ql":0.2},
                      {"f_agg":"var","isabs":"True","qh":1.0,"ql":0.8},{"f_agg":"var","isabs":"False","qh":0.6,"ql":0.4},{"f_agg":"var","isabs":"False","qh":0.8,"ql":0.4},{"f_agg":"var","isabs":"False","qh":0.4,"ql":0.2},
                      {"f_agg":"var","isabs":"True","qh":0.4,"ql":0.2},{"f_agg":"var","isabs":"True","qh":0.8,"ql":0.6},{"f_agg":"var","isabs":"True","qh":0.4,"ql":0.0},{"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.0},
                      {"f_agg":"mean","isabs":"False","qh":0.6,"ql":0.0},{"f_agg":"var","isabs":"True","qh":0.8,"ql":0.4},{"f_agg":"var","isabs":"True","qh":0.8,"ql":0.2},{"f_agg":"var","isabs":"False","qh":1.0,"ql":0.4},
                      {"f_agg":"var","isabs":"False","qh":1.0,"ql":0.8},{"f_agg":"mean","isabs":"False","qh":0.8,"ql":0.4},{"f_agg":"mean","isabs":"True","qh":0.6,"ql":0.4},{"f_agg":"mean","isabs":"True","qh":0.4,"ql":0.2},
                      {"f_agg":"mean","isabs":"True","qh":0.4,"ql":0.0},{"f_agg":"mean","isabs":"True","qh":0.2,"ql":0.0},{"f_agg":"mean","isabs":"False","qh":1.0,"ql":0.6},{"f_agg":"mean","isabs":"True","qh":0.6,"ql":0.2},
                      {"f_agg":"var","isabs":"False","qh":0.2,"ql":0.0},{"f_agg":"mean","isabs":"False","qh":0.8,"ql":0.0},{"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.6},{"f_agg":"var","isabs":"False","qh":0.8,"ql":0.0},
                      {"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.2},{"f_agg":"mean","isabs":"True","qh":0.8,"ql":0.4},{"f_agg":"var","isabs":"True","qh":1.0,"ql":0.0},{"f_agg":"var","isabs":"False","qh":1.0,"ql":0.0},
                      {"f_agg":"mean","isabs":"False","qh":0.2,"ql":0.0},{"f_agg":"var","isabs":"False","qh":1.0,"ql":0.2},{"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.2},{"f_agg":"var","isabs":"True","qh":0.8,"ql":0.0},
                      {"f_agg":"mean","isabs":"False","qh":1.0,"ql":0.8},{"f_agg":"var","isabs":"True","qh":0.6,"ql":0.4},{"f_agg":"mean","isabs":"False","qh":1.0,"ql":0.4},{"f_agg":"mean","isabs":"True","qh":0.6,"ql":0.0},
                      {"f_agg":"var","isabs":"True","qh":0.6,"ql":0.0},{"f_agg":"var","isabs":"False","qh":0.4,"ql":0.0},{"f_agg":"mean","isabs":"False","qh":1.0,"ql":0.0},
                      {"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.8},{"f_agg":"var","isabs":"True","qh":0.6,"ql":0.2},{"f_agg":"mean", "isabs":"True","qh":0.8,"ql":0.2},
                      {"f_agg":"var","isabs":"True","qh":1.0,"ql":0.2},{"f_agg":"mean","isabs":"False","qh":0.4,"ql":0.2},{"f_agg":"var","isabs":"False","qh":0.6,"ql":0.0},
                      {"f_agg":"var","isabs":"True","qh":1.0,"ql":0.6},{"f_agg":"mean","isabs":"False","qh":1.0,"ql":0.2},{"f_agg":"mean","isabs":"False","qh":0.6,"ql":0.2},
                      {"f_agg":"mean","isabs":"False","qh":0.4,"ql":0.0},{"f_agg":"mean","isabs":"True","qh":1.0,"ql":0.4},{"f_agg":"var","isabs":"True","qh":0.2,"ql":0.0}],

  "kurtosis":None
}

code:

kind_to_fc_parameters = features.json
extract_features(data,column_id = "ID",kind_to_fc_parameters=kind_to_fc_parameters)

We tried using EfficientFCParameters but it generates a lot of features that are not required.
How do we extract the features present in the JSON using extract_features()?

[kempa-liehr]

Hi @bhavara23,
kind_to_fc_parameters maps features to specific time-series. Try default_fc_parameters instead.
Cheers,
Andreas

https://tsfresh.readthedocs.io/en/latest/text/feature_extraction_settings.html