k-means 算法接受参数 k ；然后将事先输入的n个数据对象划分为 k个聚类以便使得所获得的聚类满足：同一聚类中的对象相似度较高；而不同聚类中的对象相似度较小。聚类相似度是利用各聚类中对象的均值所获得一个“中心对象”（引力中心）来进行计算的。-k-means algorithm accepts parameters k　n and the previously input data is divided into k-clustering objects in order to make the obtained cluster met: the same high similarity clustering objects　objects and different clustering Similarity small. The use of the cluster similarity clustering objects obtained by a mean of " central object"　(center of gravity) to be calculated for.