聚类算法中的k-medoids算法，和 k-means 肯定是非常相似的。事实也确实如此，k-medoids 可以算是 k-means 的一个变种。k-medoids 和 k-means 不一样的地方在于中心点的选取，在 k-medoids 算法中，我们将从当前 cluster 中选取这样一个点——它到其他所有（当前 cluster 中的）点的距离之和最小——作为中心点。-Clustering algorithm k-medoids algorithm, and k-means is certainly very similar. The fact is, k-medoids can be regarded as a variant of k-means. k-medoids and k-means not the same place that the center of the selection, in the k-medoids algorithm, we will　the current cluster in a point- its distance to all other (current cluster of) the point of and the minimum- as a central point.