Particle Swarm Optimization based Feature Selection

Abstract

Feature Selection is a pre-processing step in knowledge discovery from data (KDD) which aims at retrieving relevant data from the database beforehand. It imparts quality to the results of data mining tasks by selecting optimal feature set from larger set of features. Various feature selection techniques have been proposed in past which, unfortunately, suffer from unavoidable problems such as high computational cost and getting stuck into the local optima. Evolutionary algorithms such as Particle Swarm Optimization (PSO) possess immense abilities to explore a large search space and rarely fall into local optima thus making them a nice choice for feature selection. In this paper, we have explored pros and cons of traditional and PSO based feature selection techniques and suggested some effective changes in existing approaches.

Sleep stage classification using extreme learning machine and particle swarm optimization for healthcare big data

“…(a) Randomly initializing velocity and particle population; (b) Starting new iteration; (c) Evaluating the fitness function for all particles; (d) Determining p-best of each particle and replacing the p-best when it is better than the previous one; (e) Determining g-best value; (f ) Updating each particle velocity by Eq. (15).…”

“…To aid the feature selection process and determining hidden node number, PSO is chosen as algorithm in this work. Previous study has shown a promising result in using PSO for selecting features [15]. PSO showed minor accuracy reduction with significantly faster performance compared to other algorithms such as GA [16].…”