A New Interval Type-2 Fuzzy PRISM Algorithm

Interval type-2 rule-based fuzzy systems become popular in the last years. However, the complexity of operations on an interval membership values cause that there are only a few methods that allow inducing type-2 fuzzy rules from data. Moreover, the existing algorithms most often rely on the induction of decision trees and can lead to an overly large rule base. For those reasons, the form of fuzzy rules is most often determined by an expert. In this paper, we would like to introduce a new rule induction algorithm based on the PRISM strategy. As the exemplary application of the proposed method show, it allows obtaining the rule base that is more compact and more manageable.