High Speed Training of a Fuzzy Classifier with Ellipsoidal Regions

Abstract

In this paper, we discuss a fast training method for a fuzzy classifier with ellipsoidal regions. In the calculation of the degree of membership, first we decompose the covariance matrix into triangular matrices. Then we set small offdiagonal elements relative to diagonal elements to zero, and calculate the degree of membership by skipping zero elements. We demonstrate the effectiveness of our method using several benchmark data.