Salp Swarm Algorithm based Semi-supervised Metric Fuzzy Clustering

Abstract

Semi-supervised Metric Fuzzy Clustering (SMUC) is known for taking advantage of prior information of membership to guide clustering. However, SMUC has the following problem: it is very easy to lead to lower utilization effect of prior information of membership due to sensitivity to the initial cluster centers. In order to solve the problem, we propose a hybrid approach for Salp Swarm Algorithm based Semi-supervised Metric Fuzzy Clustering (SSA-SMUC). Firstly, several data samples were randomly selected as the salp population from all data samples to obtain the optimal initial cluster centers. Secondly, the optimal initial cluster centers are obtained by the Salp Swarm Algorithm using the DB index as its fitness function. Finally, applying the optimal cluster centers as the initial cluster centers of SMUC for clustering, we can improve SMUC's utilization effect of prior information and get the better clustering result. Further experiments conducted in UCI data show that the proposed clustering algorithm can derive a better performance.