A Multiobjective Genetic Approach with TOPSIS to Learn Fuzzy Rule-Based Classification Systems

The design of a Fuzzy Rule-Based Classification System (FRBCS) taking into account accuracy and interpretability (number of rules) of the FRBCS is an important issue. In this paper, we extend our previous work by considering the average number of fuzzy labels per input variable as a new objective to be treated with TOPSIS. In addition, we employ information entropy to determine the weights of the evaluation criteria. We compare our proposal with a method that deals with the accuracy-interpretability tradeoff by means of a multiobjective evolutionary process based on the Non-dominated Sorting Genetic Algorithm II (NSGA-II). The obtained results show the effectiveness of the proposed approach.