Solving Feature Selection Problems Built on Population‐based Metaheuristic Algorithms
Mohamed Sassi · 2022
This chapter demonstrates the use of the grey wolf optimization for the feature selection process applied on a k-nearest neighbor classifier with very competitive results compared to the genetic algorithm and Particle Swarm Optimization metaheuristics. The keystone of the grey wolf pack is its hierarchy, which is structured into four groups: Alpha, Beta, Delta and Omega. The grey wolf pack hunts their prey by progressing through four phases: search, pursue, surround and harass the prey until trapped or exhausted, and attack the prey. The chapter discusses the mathematical modeling of optimization in a binary discrete search space and provides the binarization modules allowing continuous metaheuristics to be able to solve feature selection problems in binary search space. The binary feature selection family has two predominant types of feature selection methods: filter method and wrapper method. Feature selection provides strategic advantages in the training of classifiers by increasing their performance while reducing their complexity.