一种基于K-means的分布式聚类算法

A distributed clustering algorithm based on K-means algorithm

为解决现有的分布式聚类算法效率低下和不能保护数据隐私的问题,在K-Dmeans算法的基础上,提出一种新的分布式聚类算法。该算法利用数据对象间的密度函数值来优化站点初始聚类中心,从而大大降低了聚类的迭代次数;同时各从站点只需向主站点传送其聚簇的特征信息,有效降低分布式聚类过程中的通信量,保护了各个站点的独立性。实验结果表明,该算法有效可行,且在效率和聚类质量上优于K-Dmeans。

In order to improve the privacy protection and efficiency of the existing distributed clustering algorithms.A new algorithm based on the K-Dmeans algorithm was proposed,The algorithm used the value of the density function to optimize the sets＇ centers of K-means clustering algorithm,significantly reducing the number of iterations of clustering,while the slaver sets only sent the information of the clustering,reducing the distributed clustering process of traffic and protecting the independence of each site.Experimental results show that the proposed clustering algorithm is effective and feasible,which is better than K-Dmeans and K-means algorithm in efficiency and quality.