摘要
针对K-均值算法易受孤立点影响、对初始中心点选择敏感、易陷入局部最优的问题,对K-均值算法进行了改进,提出了一种自适应优化选择初始中心点的K-均值算法。实验结果表明,改进后的算法不仅较大程度上弥补了传统K-均值算法的不足,并且提高了聚类的稳定性和准确率。
For the problems that k- means algorithm is susceptible to outlier influence,sensitive to the choice of initial center, easy to fall into local optimum, an adaptive optimization method to select the initial center is proposed to improve k- means algorithm. Experimental results show that the improved algorithm can compensate for the lack of k- means algorithm in a large extent,and owns higher stability and greater accuracy.
出处
《济源职业技术学院学报》
2014年第4期4-7,共4页
Journal of Jiyuan Vocational and Technical College
关键词
K-均值
聚类算法
数据挖掘
K-means algorithm
cluster algorithm
data mining