实现K均值聚类，N为分类个数，data作为输入数据，re为输出带标签数据，u是最终聚类中心(To achieve K means clustering, N is the number of classification, data as input data, re as output tag data, and u is the final clustering center)