A Survey on Interpretability-Accuracy (I-A) Trade-Off in Evolutionary Fuzzy Systems

Obtaining high degree of Interpretability and Accuracy in the design of complex high dimensional fuzzy systems is contradictory in nature. It is well known by Interpretability-Accuracy (I-A) Trade-Off. Several methods have been proposed to find good I-A Trade-Off. These includes Multi objective Optimization, Context Adaptation, Hierarchical Fuzzy Modeling and other issues related to fuzzy partition, membership functions, rules, rule bases etc. This paper is the survey of different approaches developed regarding the issue of I-A Trade-Off in designing fuzzy systems in the last decade. Also, the techniques for assessing the interpretability and accuracy have been surveyed.