期刊名称：International Journal on Computer Science and Engineering
印刷版ISSN：2229-5631
电子版ISSN：0975-3397
出版年度：2012
卷号：4
期号：06
页码：1174-1180
出版社：Engg Journals Publications
摘要：In k-means clustering algorithm, the number of centroids is equal to the number of the clusters in which data has to be partitioned which in turn is taken as an input parameter. The initial centroids in original k-means are chosen randomly from the given dataset and for the same dataset different clustering results are produced with different randomly chosen initial centroids. This paper presents a solution to this limitation of the original K-means Algorithm.
