Voting Schemes in Fuzzy Classification Systems

Abstract

In this paper, we examine two kinds of voting schemes in fuzzy-rule-based classification systems. One is based on the voting by multiple fuzzy if-then rules in a single fuzzy classification system. The other is based on the voting by multiple fuzzy classification systems. First we describe the voting by multiple fuzzy if-then rules, which is used for the fuzzy reasoning in a single fuzzy-rule-based classification system. We also explain how the grade of certainty of each fuzzy if-then rule can be adjusted when the voting by fuzzy if-then rules is used for the fuzzy reasoning. Then we describe three voting methods for combining the classification results of multiple fuzzy classification systems : a perfect unison rule, a majority rule and a weighted voting rule. By computer simulations on real-world pattern classification problems such as iris data, appendicitis date and cancer date, we examine the performance of these voting methods. Simulation results show that the performance of the voting by multiple fuzzy classification systems is better than that of each individual classification system. The performance of the voting by multiple fuzzy classification systems is also compared with various results of other classification methods reported in the literature.