Ensemble? and distance?based feature ranking for unsupervised learning

In this work, we propose two novel (groups of) methods for unsupervised feature ranking and selection. The first group includes scores (Genie3 score, RandomForest score) that are computed from ensembles of predictive clustering trees. second method is URelief, the extension Relief family algorithms. Using 26 benchmark data sets 5 baselines, show both Genie3 score (computed ensemble extra trees) URelief outperform existing performs best overall, in terms power top-ranked features. Additionally, analyze influence hyper-parameters proposed on their performance, highest quality achieved by most efficient parameter configuration. Finally, a way discovering location features ranking, which relevant reality.