k均值聚类算法，使各个样本与所在类均值的误差平方和达到最小，并且附有显示程序-k-means clustering algorithm, where the class so that each sample and the mean squared error to a minimum, and with the display program