A Review of Interval Valued Type 2 Fuzzy Rule-Based Classifiers
T. Warren Liao · 2023
This paper presents a review of 21 past works on developing interval-valued type-2 fuzzy rule-based classifiers (IVT2FRBC) for various applications. Special attention is paid to two major topics: design and performance evaluation. The first topic involves decisions made pertaining to every relevant component of IVT2FRBC such as rule structure, number of input variables/features, type & number of membership functions for each variable, model construction method, model tuning, and so on. The second topic includes measures used to evaluate a classifier's performance, classifiers compared, and statistical tests carried out to describe the performance differences between/among classifiers. Based on the review, observations are made as to their soundness and usefulness.