Prototypes-Based Horizontal Collaborative Fuzzy Clustering

Abstract

As a typical extension of fuzzy c-means, horizontal collaborative fuzzy clustering implements the clustering on a data set of some patterns with the collaboration of some knowledge which is obtained from other data set(s) about the same patterns but described in different feature space(s). For the sake of safety and personal privacy, the knowledge is usually provided by partition matrixes. This paper presents a new approach to implementing horizontal collaborative fuzzy clustering with the knowledge provided by the prototypes instead of partition matrixes. This prototypes-based horizontal collaborative fuzzy clustering algorithm exhibits good performance which is showed in our experiments.