A new method for constructing membership functions and fuzzy rulesfrom training examples

摘要

To extract knowledge from a set of numerical data and build up anrule-based system is an important research topic in knowledgenacquisition and expert systems. In recent years, many fuzzy systems thatnautomatically generate fuzzy rules from numerical data have beennproposed. In this paper, we propose a new fuzzy learning algorithm basednon the Α-cuts of equivalence relations and the Α-cuts ofnfuzzy sets to construct the membership functions of the input variablesnand the output variables of fuzzy rules and to induce the fuzzy rulesnfrom the numerical training data set. Based on the proposed fuzzynlearning algorithm, we also implemented a program on a Pentium PC usingnthe MATLAB development tool to deal with the Iris data classificationnproblem. The experimental results show that the proposed fuzzy learningnalgorithm has a higher average classification ratio and can generatenfewer rules than the existing algorithm