首先从n个数据对象任意选择 k 个对象作为初始聚类中心；而对于所剩下其它对象，则根据它们与这些聚类中心的相似度（距离），分别将它们分配给与其最相似的（聚类中心所代表的）聚类；然后再计算每个所获新聚类的聚类中心（该聚类中所有对象的均值）；不断重复这一过程直到标准测度函数开始收敛为止。一般都采用均方差作为标准测度函数. k个聚类具有以下特点：各聚类本身尽可能的紧凑，而各聚类之间尽可能的分开。-First, a data object from the n choose k objects as initial cluster centers　and for the rest of the other objects, according to their similarity with the cluster center (distance), respectively, assign them to their most similar (represented by cluster center) clustering　then calculated for each cluster center received a new clustering (all objects in the cluster mean)　repeats this process until the convergence criteria begin until the measure function. Standard deviation is generally used as a standard measure function. K a cluster has the following characteristics: the cluster itself as a compact, but separated as much as possible between each cluster.