Sequential Feature Classification in the Context of Redundancies

TL;DR

This paper introduces a novel method using random forests and statistical techniques to distinguish between strong and weak relevance in feature selection for non-linear problems, addressing a gap in existing linear-only approaches.

Contribution

It extends relevance distinction methods from linear to non-linear problems using random forest models and statistical analysis.

Findings

Successfully differentiates strong and weak relevance in non-linear feature selection

Adapts relevance distinction to non-linear models using random forests

Provides a new approach applicable beyond linear problem limitations

Abstract

The problem of all-relevant feature selection is concerned with finding a relevant feature set with preserved redundancies. There exist several approximations to solve this problem but only one could give a distinction between strong and weak relevance. This approach was limited to the case of linear problems. In this work, we present a new solution for this distinction in the non-linear case through the use of random forest models and statistical methods.