Processes

Causal Analysis

class hawk.processes.wps_causal.Causal[source]

causal Causal Analysis (v1.5)

Performs a causal analysis with multiple configurations, returning outputs plots and pkl file.

Parameters:

  • dataset_train (text/csv) – Please add the train csv file here.

  • dataset_test (text/csv) – Please add the test csv file here.

  • target_column_name (string) – Please enter the case-specific name of the target variable in the dataframe.

  • pcmci_test_choice ({'ParCorr', 'CMIknn'}) – Choose the independence test to be used in PCMCI.

  • pcmci_max_lag ({'0', '1', '2', '3', '4', '5'}) – Choose the maximum lag to test used in PCMCI.

  • tefs_direction ({'forward', 'backward', 'both'}) – Choose the direction of the TEFS algorithm.

  • tefs_use_contemporary_features (boolean) – Choose whether to use contemporary features in the TEFS algorithm.

  • tefs_max_lag_features ({'no_lag', '1', '2', '3', '4', '5'}) – Choose the maximum lag of the features in the TEFS algorithm.

  • tefs_max_lag_target ({'1', '2', '3', '4', '5'}) – Choose the maximum lag of the target in the TEFS algorithm.

Returns:

  • pkl_baseline (application/octet-stream) – The baseline scores on the initial data.

  • plot_pcmci (application/pdf) – The selected features by PCMCI.

  • plot_pcmci_preview (image/png) – The selected features by PCMCI.

  • pkl_pcmci (application/octet-stream) – The PCMCI results details.

  • plot_tefs (application/pdf) – The selected features by TEFS.

  • plot_tefs_preview (image/png) – The selected features by TEFS.

  • pkl_tefs (application/octet-stream) – The TEFS results.

  • plot_tefs_wrapper (application/pdf) – The wrapper scores evolution by TEFS.

  • plot_tefs_wrapper_preview (image/png) – The wrapper scores evolution by TEFS.

  • pkl_tefs_wrapper (application/octet-stream) – The TEFS wrapper scores evolution details.

References