Constructing fuzzy models with linguistic integrity from numerical data-AFRELI algorithm

This paper presents an algorithm to extract rules relating input/output data and including prior knowledge. The rules are created in the environment of fuzzy systems. The fuzzy sets describing the system are constructed within a framework of linguistic integrity to guarantee its interpretability in the linguistic context. Two algorithms are presented in this paper. The main algorithm is the autonomous fuzzy rule extractor with linguistic integrity (AFRELI). This algorithm is complemented with the use of the FuZion algorithm created to merge consecutive membership functions, while guaranteeing the distinguish ability between fuzzy sets. Comparisons with other proposed methods show a good tradeoff between accuracy and interpretability.