The initial clustering center of the traditional K-means algorithm was generated randomly from the data set, and the clustering result was unstable.

传统的K均值算法的初始聚类中心从数据集中随机产生，聚类结果很不稳定。