A new hybrid model of PSO and DE algorithm for data classification

Abstract

This paper presents a new hybrid HPSO-DE classification algorithm that combines the advantages of particle swarm optimization algorithm and differential evolution algorithm. Major improvements achieved by this combination are 1) flight improvement — flight behaviors are more and better diversified because each of the top 3 particles gets put into 3 different groups of the rest and then each group is mutated with a different operator and 2) particle improvement — members of a succeeding generation are composed of more of better particles than those of the current generation because better particles are allowed to produce more offspring. HPSO-DE and several other classification models were performance tested with 8 benchmarking datasets, and HPSO-DE was found to outperform them on 6 out of the 8.