K-mw-modes: An algorithm for clustering categorical matrix-object data

Abstract

In data mining, the input of most algorithms is a set of n objects and each object is described by a feature vector. However, in many real database applications, an object is described by more than one feature vector. In this paper, we call an object described by more than one feature vector as a matrix-object and a data set consisting of matrix-objects as a matrix-object data set. We propose a k-multi-weighted-modes (abbr. k-mw-modes) algorithm for clustering categorical matrix-object data. In this algorithm, we define the distance between two categorical matrix-objects and a multi-weighted-modes representation of cluster prototypes is proposed. We give a heuristic method to choose the locally optimal multi-weighted-modes in the iteration of the k-mw-modes algorithm. We validated the effectiveness and benefits of the k-mw-modes algorithm on the five real data sets from different applications.