Hierarchical fuzzy model for classification problem

This paper proposes a new methodology to solve the classification problem, where the fuzzy IF-THEN rule with hierarchy framework is utilized. The number of rules and the correct classification rate are the essential requirements for classification problem. The proposed scheme can acquire higher classification rate with few fuzzy partitions and fuzzy rules. The developed model comprises two stages; one is generation of fuzzy IF-THEN rules for the subsystems and the other is to determine the decision unit. The performance has been tested by computer simulations on the well-known Wine and Iris databases. Simulations demonstrate that our method under a few rules can provide sufficiently high classification rate even with higher feature dimension.