A Hybrid Approach for Optimal Feature Selection based on Evolutionary Algorithms and Classic Approaches

Abstract

Feature selection (FS) is a fundamental problem in the field of pattern recognition, which aims to find a minimal feature subset from the original feature space while retaining a suitably high accuracy in representing the original features. FS is used to improve the efficiency of learning algorithm especially for large scale datasets, by finding a minimal subset of features that has maximum efficacy on classifier. In this paper, we proposed a new hybrid approach based on Evolutionary Algorithms and Heuristic methods for effective feature selection. In other words, the proposed approach has a hybrid heuristic/random strategy for search optimal solution. We compare the obtained simulation results with other algorithms separately, like evolutionary algorithms (with the same situation like iteration, population and cost function) consist on genetic algorithm (GA), ant colony optimization (ACO) and particle swarm optimization (PSO), and also with Heuristic Methods consist on sequential forward selection (SFS) and sequential backward elimination (SBE). Obtained results demonstrate that the proposed hybrid algorithm is effective and efficient for effective feature selection.