Genetic Interval Type-2 Fuzzy Classifier Generation: A Comparative Approach

This paper concentrates on studying the use of interval type-2 fuzzy sets for the pattern classification problem. Even though researchers recognize that type-2 fuzzy sets are more difficult to understand and use than type-1 fuzzy sets, the interest in the study is motivated by the additional power to represent uncertainty in different levels. The work developed here relies on the recent advances concerning the three-dimensional type-2 membership functions to focus on the genetic generation of type-2 fuzzy classifiers. We use a three stage Genetic Algorithm Architecture to generate Fuzzy Classification Systems, composed of three Genetic Algorithms that generate the rule base, optimize the interval type-2 membership functions and optimize the number of rules. With the objective of contributing to the discussion concerning the benefits and costs of using type-2 fuzzy sets, this paper presents additional experiments and analysis, using datasets from the UCI Machine Learning Repository. Fuzzy classifiers were generated using the Genetic Algorithm Architecture for both type-1 and type-2 fuzzy sets, and using another genetic generation method found in the literature. The results demonstrated that the type-2 fuzzy classifier presents better performance with a small number of rules.