Improving feature selection algorithms using normalised feature histograms

TL;DR

This paper introduces a feature selection method that uses normalized feature histograms to improve stability and accuracy across diverse datasets, outperforming traditional methods.

Contribution

The paper presents a novel histogram-based feature selection approach that enhances stability and classification performance over conventional techniques.

Findings

Significant accuracy improvements on microarray and image datasets.

Reduced feature instability compared to traditional methods.

Effective across multiple datasets and feature selection criteria.

Abstract

The proposed feature selection method builds a histogram of the most stable features from random subsets of a training set and ranks the features based on a classifier based cross-validation. This approach reduces the instability of features obtained by conventional feature selection methods that occur with variation in training data and selection criteria. Classification results on four microarray and three image datasets using three major feature selection criteria and a naive Bayes classifier show considerable improvement over benchmark results.