摘要
通过分析传统K-means聚类算法初始化随机和聚类结果难以体现对象间相关性的问题,提出了一种基于相关性加权改进的K-means算法.该算法通过引入能够反映对象间相关性程度的权衡因子皮尔逊相关性系数,利用经归一化后的相关性系数对欧式距离进行加权,对传统的K-means算法进行改进.实验结果表明:文中改进后的算法相比传统K-means算法和其它改进算法,在聚类质量上能获得更佳的聚类效果.
In view of the fact that it is difficult to reflect the inter-object correlation problem by analyzing the random initialization and clustering results of the traditional k-means, an improved k-means algorithm based on correlation weighting is proposed. Firstly, it can reflect the degree of correlation between the objects by introducing the Pearson correlation coefficient. Secondly, the traditional k-means is improved by weighting the Euclidean distance by using normalized correlation coefficients. Experimental results show that the proposed algorithm can obtain better effects on the clustering quality compared with the traditional k-means algorithms and other improved ones.
出处
《江西理工大学学报》
CAS
2018年第1期87-92,共6页
Journal of Jiangxi University of Science and Technology
基金
国家自然科学基金资助项目(61462036)
