基于Kd树改进的高效K-means聚类算法

摘要

This paper proposed a modified clustering algorithm called ck-means for the question that the classic k-means algorithm is expected to be improved in the clustering efficiency of multidimensional data.Based on the k-means algorithm, the pk-means algorithm introduces a spatial data structure called kd tree,selects the initial clustering center points from the equal-interval partition point sets of a certain dimension of the multidimensional data,and adopts pruning strategy during the points assigning process,to improve the clustering efficiency of the algorithm.Experimental results show that the clustering efficiency of ck-means is higher than that of the k-means .%针对经典的K-means算法在多维数据聚类效率上还有待提高的问题,本文提出一种称为 CK-means的改进聚类算法。该算法在 k-means算法的基础上,通过引入 Kd 树空间数据结构,初始聚类中心从多维数据某一维的区间等间隔集中选取,以及在数据对象分配过程中采用剪枝策略来提高算法的运行效率。实验结果表明,CK-means聚类算法较经典的 k-means聚类算法运行效率更高。