A Wolf Pack Optimization Theory Based Improved Density Peaks Clustering Approach

Abstract

In view of the problem that the Density Peaks Clustering (DPC) algorithm needs to manually set the parameter cut-off distance (dc) we propose a Wolf Pack optimization theory based Density Peaks Clustering approach (WPA-DPC). Firstly, we introduce dc parameter into the Wolf Pack Algorithm (WPA) to speed up the search. Secondly, we introduce the WPA into the DPC algorithm; the cut-off distance is used as the location of the wolf group. Finally, we make silhouette index in the search process as the fitness value, and the optimal location of the wolf group is the parameter value at the end. The simulation results show that compared with the traditional Density Peaks Clustering algorithm, the proposed algorithm is closer to the true clustering number. According to the evaluation results of silhouette and f-measure, the quality of clustering and the accuracy are greatly improved.