Graded Possibilistic Approach to Variable Selection in Linear Fuzzy Clustering

Linear fuzzy clustering is a useful tool for knowledge discovery in databases (KDD), and several modifications have been proposed in order to analyze real world data. This paper proposes a new approach for estimation of local linear models, in which linear fuzzy clustering is performed by selecting variables in each cluster. The new clustering model uses two types of memberships. One is the conventional membership that represents the degree of membership of each sample. The other is the additional parameter that represents the responsibility of each variable and is given by the graded possibilistic approach. Numerical experiments demonstrate the characteristics of the proposed technique