A classification algorithm based on simplified fuzzy rules base

Abstract

This paper proposes a classification algorithm based on simplified fuzzy rules base combining fuzzy clustering with rough set. Firstly, generates fuzzy rules base using fuzzy clustering from numerical sample dates, and then simplifies the sample attributions using rough set theory, deletes the redundant rules, and gets the simplified fuzzy rules base, in order to make classification decision conveniently. The performance of the classification algorithm is tested by the IRIS data, and the results show that the fuzzy rules are not only intelligible, but also have very good classification performance.