Enhancing Gene Selection in Microarray Dataset Using Binary Gray Wolf Optimization Algorithm and Statistical Dependence Technique

Abstract

The problem of selecting the most relevant subset of features is vital for enhancing classification accuracy while minimizing computational load in machine learning. To tackle this challenge, the paper investigates two approaches: the Binary Gray Wolf Optimization (BGWO) algorithm and the Statistical Dependence (SD) technique. The process begins with the SD technique to determine the features that most significantly influence classification outcomes. Then, the BGWO algorithm is applied in conjunction with the K-Nearest Neighbors (KNN) classifier to further narrow down the selection to the most essential features. The proposed SD-BGWO approach outperforms conventional methods by either improving classification accuracy or by reducing the number of features required, thereby optimizing the feature selection process in terms of both efficiency and effectiveness.