Rough Set Attribute Reduction Based on Genetic Algorithm

In order to overcome the difficulties in attribute reduction with large quantity of condition attributes, genetic algorithm was employed to obtain the minimal reduction of decision tables under existed conditions by combining its outstanding ability for overall searching with rough set theory. A fitness function was proposed and applied to the genetic algorithm, which accelerated the speed of convergence. The detailed algorithm and the computation process were presented for practical purpose. The simulation results show that the proposed approach has good searching ability and high restraining speed and can achieve efficient attribute reduction.