计算机应用与软件2017，Vol.34Issue(3)：212-217,6.DOI:10.3969/j.issn.1000-386x.2017.03.038
密度峰值优化初始中心的K-means算法
K-MEANS ALGORITHM OF OPTIMIZED INITIAL CENTER BY DENSITY PEAKS
摘要
Abstract
K-means algorithm randomly selects the initial cluster centers, which can easily lead to the instability of clustering results.To overcome this deficiency, a K-means algorithm named clustering by fast search and find of density peaks (CFSFDP) is proposed to optimize the initial center.Firstly, aiming at the disadvantage that the selection of truncated distance in CFSFDP algorithm affects the local density, it is proposed that the local density of each point can be replaced by gravitational potential energy.Based on this, K-means clustering can be implemented by using the optimized CFSFDP algorithm to select initial cluster centers.The proposed algorithm is tested on UCI data sets and synthetic data sets.Experimental results show that the new algorithm can achieve better clustering.关键词
K-means算法/CFSFDP算法/密度峰值/引力势能Key words
K-means algorithm/CFSFDP algorithm/Density peak/Gravitational potential energy分类
信息技术与安全科学引用本文复制引用
李敏,张桂珠..密度峰值优化初始中心的K-means算法[J].计算机应用与软件,2017,34(3):212-217,6.基金项目
江苏省自然科学基金项目(BK20140165). （BK20140165）
