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

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 highaccuracy in representing the original features. FS is used to improve the efficiency of learning algorithm especially for largescale datasets, by finding a minimal subset of features that has maximum efficacy on classifier In this paper, we proposed a new hybrid approach based onEvolutionary Algorithms and Heuristic methods for effective feature selection. In other words, the proposed approach has ahybrid heuristic/random strategy for search optimal solution. We compare the obtained simulation results with other algorithmsseparately, 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 sequentialbackward elimination (SBE). Obtained results demonstrate that the proposed hybrid algorithm is effective and efficient for effective feature selection

کلیدواژه ها:

Feature selection ، SFS ، SBE ، optimization