一种自动获得k值的聚类算法

A Getting k Automatically Based Clustering Algorithm

针对k均值算法局限于k值和初始中心点选取的情况,提出了一种基于k均值的自动获得k值的KDM算法.该算法整体沿用k均值算法的思想,利用最大最小距离法选择初始聚类中心,并且选择聚类中心与划分对象同时进行.通过不断改变类中心,来达到较好的聚类效果.

Against the defect of k -means, a new algorithm called KDM is proposed based on k - means and DIANA. According to DIANA, the method of max-min distance is used to choose the initial center, and cluster the data at the same time, and better result is achieved by changing the center.