(This is an old code which I've been using to select Best candidate from a cluster).
Major changes
Implements bps_numerical.feature_selection.BestCandidateFeatureSelector based on clusters formed
Implements shap-based feature scorer at bps_numerical.classification.feature_scorers.ShapBasedPhenotypeFeatureScorer
Minor changes
Improvements of plotting tools at misc
Addition of misc.datatools.LoadSaveMixin for quick load/save of custom components. This mixin can now enable the saving & loading of clustering, feature selection, and gene ranking components.
Speeds up CorrelationClusterer clustering component by caching the distance matrix (can be provided externally through the constructor)
(This is an old code which I've been using to select
Best
candidate from a cluster).Major changes
bps_numerical.feature_selection.BestCandidateFeatureSelector
based on clusters formedbps_numerical.classification.feature_scorers.ShapBasedPhenotypeFeatureScorer
Minor changes
misc
misc.datatools.LoadSaveMixin
for quick load/save of custom components. This mixin can now enable the saving & loading of clustering, feature selection, and gene ranking components.CorrelationClusterer
clustering component by caching the distance matrix (can be provided externally through the constructor)