Get guide features importance

Returns permutation-based (as in ranger) feature importance of features used to guide the clustering, for one given cluster.

get_feature_importance(stoic_results, cluster)

Arguments

stoic_results

Output object of `stoic_clustering`.

cluster

Number of the cluster of interest (numeric).

Value

A dataframe of features ranked by importance.

Details

The direction column is TRUE when the feature has higher values in the cluster than in other observations, FALSE otherwise. The former is more easily interpreted as it corresponds to a homogeneous set of observations.