关键词: K_means聚类, 初始聚类中心, 启发式信息, 收敛速度

Abstract: This paper presents an improved clustering model based on distance measurement, in order to solve the problem of slow convergence rate of traditional K_means clustering method by selecting initial cluster centers randomly. By using effective heuristic information, this method selects better clustering centers and reduces the iteration steps of attaining stable clustering state. Then the speed of algorithm is accelerated. Simulation results on UCI datasets demonstrate that comparing with traditional K_means clustering means, the improved K_means has fast convergence rate and the better clustering results are obtained by this model after less iterations.

Key words: K_means clustering, initial clustering center, heuristic information, convergence rate