Fuzzy decision trees by fuzzy ID3 algorithm and its application to diagnosis systems

Abstract

A popular and particularly efficient method for making a decision tree for classification from symbolic data is ID3 algorithm. Revised algorithms for numerical data have been proposed, some of which divide a numerical range into several intervals or fuzzy intervals. Their decision trees, however, are not easy to understand. We propose a new version of ID3 algorithm to generate an understandable fuzzy decision tree using fuzzy sets defined by a user. We apply it to diagnosis for potential transformers by analyzing gas in oil.<>