tick.preprocessing

This module provides several preprocessing utilities, in the form of transformer classes that change raw feature vectors into a suitable representation for some learners. These transformers should be scikit-learn compatible, whenever possible.

Preprocessing for static features

The FeaturesBinarizer binarizes all continuous features found in features matrix. This transformer is particularly useful whenever using the ProxBinarsity penalization for supervised linear learning see tick.linear_model.

FeaturesBinarizer([method, n_cuts, …])

Transforms continuous data into bucketed binary data.

Preprocessing for longitudinal features

This module also provides preprocessor specific to longitudinal features with a similar API to scikit-learn preprocessors.

LongitudinalFeaturesProduct([exposure_type, …])

Transforms longitudinal exposure features to add the corresponding product features.

LongitudinalFeaturesLagger(n_lags[, n_jobs])

Transforms longitudinal exposure features to add columns representing lagged features.

LongitudinalSamplesFilter([n_jobs])

Longitudinal data preprocessor which filters out samples for which all