Skip to content

v0.2.7

Latest

Choose a tag to compare

@malgar malgar released this 13 Aug 21:48

Release notes:

Clustering

  • Optimal clustering computation with other metrics than inertia uses metrics scores for clusters >1
  • Silhouette and Calinski and Harabasz scores are concave increasing (instead of convex decreasing)

Classifier

  • plot_shap_importances() : pandas.DataFrame.append was deprecated in version 1.4.0. It has been replaced by pandas.concat