A Fuzzy Approach to Robust Clusterwise Regression

Abstract

new robust fuzzy linear clustering method is proposed. We estimate coe cients of a linear regression model in each unknown cluster. Our method aims to achieve robustness by trimming a xed proportion of observations. Assignments to clusters are fuzzy: observations contribute to estimates in more than one single cluster. We describe general criteria for tuning the method. The proposed method seems to be robust with respect to di erent types of contamination.