ファジィ理論を用いた診断支援システム : 多群鑑別システムの構築(一般講演7-ファジィ推論-)

抄録

This paper introduces an algorithm of membership function acquisition in consequent clause in fuzzy diagnostic systems for multi-group differentiation. In this algorithm, the relative positions and widths of membership functions in consequent clause are obtained by calculation of the possibility of fuzzy events and the information content of fuzzy sets in antecedent clauses. Moreover, the membership functions are detected by optimization using learning data. To evaluate the performance of this algorithm, we have applied it on Fisher's iris data. As a result, the accuracy is more than 98%.