Skip to Main Content
In this paper, a new Multi-Objective Particle Swarm Optimization (MOPSO) is applied to solve a problem of feature selection defined as a multiobjective problem. This algorithm (pMOPSO), known for its fast convergence with negligible computation time is based on a distributed architecture. Sub-swarms are obtained from dynamic subdivision of the population using Pareto Fronts. The algorithm addresses a problem defined by two goals, characterized by their contradictory aspect, namely, minimizing the error rate and minimizing the number of features. The two objectives are treated simultaneously constituting the objective function. Performance of our approach is compared with other evolutionary techniques using databases choosing from the UCI repository .