Using multi-objective particle swarm optimization for designing novel classifiers

Abstract

The effectiveness and powerfulness of the multi-objective swarm intelligence algorithms are motivations for utilizing them to estimate the optimum decision functions/decision rules in such a way that various performance aspects of classifiers (e.g., score of recognition and reliability) are simultaneously optimized. This chapter explains the applications of multi-objective swarm intelligence techniques (especially particle swarm optimization) on designing novel classifiers (named direct applications) and optimizing the performance aspects of conventional classifiers (named indirect applications). Also, a review of some of the past and ongoing related research is presented. © 2009 Springer-Verlag Berlin Heidelberg.