Extracting relevant information from input-output data

Abstract

In this paper a set of heuristic criteria devised to address the problems encountered in designing a fuzzy system from input-output data is presented. In particular, we show how to discriminate unsignif­icant linguistic variables, determine the number of fuzzy sets, place them in the universe of scope, and propose a set of linguistic rules. The objective is to obtain in a simple and fast manner an algorithm simpler than the standard ones with mínimum com­putational cost but still similar performance and more intelligibility in most cases.