Rules Extraction by Clustering Artificial Fish-swarm and Rough Set

Due to the ill-conditioned problem caused by inefficient discretization approaches, it is difficult for the traditional rough set theory to extract accurate rules. And the continuous value needs to be discretized in the process of rule extraction. Then in this paper, a method based on clustering Artificial Fish-Swarm Algorithm (AFSA) and rough set theory is proposed to extract decision rules. Firstly, the clustering algorithm is used to classify attribute values in accordance with decision attributes. Secondly, the artificial fish-swarm algorithm is used to discretize the continuous attributes and to reduce the decision table. The experimental results indicate that the decision rules derived from the proposed method are much simpler and more precise.