A modified form of mutation for genetic-fuzzy classifier design

Abstract

This paper presents a Genetic Algorithm (GA) approach to obtain the optimal rule set and the membership function. While designing the fuzzy classifier using GA, the membership functions are represented as real numbers and the rule set is represented by the binary string. BLX-a crossover is used for real numbers and two point crossover and an advanced operator called gene cross swap operator are used for the binary string. A modified form of mutation that uses the concept of velocity updating in Particle Swarm Optimization (PSO) is proposed to improve the convergence speed and quality of the solution. The performance of the proposed approach is evaluated through development of fuzzy classifier for four standard data sets. Simulation results show that the proposed algorithm produces a fuzzy classifier with minimum number of rules and high classification accuracy.