Traditional K-Means algorithm is sensitive to the initial centers and easy to get stuck at locally optimal value.

传统K均值算法对初始聚类中心敏感，聚类结果随不同的初始输入而波动，容易陷入局部最优值。