Confident Feature Ranking

TL;DR

This paper introduces a framework and a novel method for quantifying uncertainty in feature importance rankings, providing confidence intervals for feature ranks and improving interpretability of model explanations.

Contribution

It presents a new approach for post-hoc feature importance interpretation that accounts for uncertainty and offers confidence intervals for feature ranks.

Findings

Provides simultaneous confidence intervals for feature ranks

Enables reliable selection of top-k important features

Improves stability of feature importance interpretation

Abstract

Machine learning models are widely applied in various fields. Stakeholders often use post-hoc feature importance methods to better understand the input features' contribution to the models' predictions. The interpretation of the importance values provided by these methods is frequently based on the relative order of the features (their ranking) rather than the importance values themselves. Since the order may be unstable, we present a framework for quantifying the uncertainty in global importance values. We propose a novel method for the post-hoc interpretation of feature importance values that is based on the framework and pairwise comparisons of the feature importance values. This method produces simultaneous confidence intervals for the features' ranks, which include the ``true'' (infinite sample) ranks with high probability, and enables the selection of the set of the top-k important features.