基于k-means的(1+ε)近似算法求解(1+ε)-Approximation Algorithm for k-means
王守强,朱大铭
摘要(Abstract):
针对Amit Kumar提出的求解k-means算法的1+ε近似求解随机算法,提出了一个改进措施用来提高每次取样的成功概率。当固定k和ε值时该算法为线性的。通过多次运行该算法能以较高的概率求出k-means算法的1+ε近似值。
关键词(KeyWords): k-means;聚类;随机算法;取样
基金项目(Foundation):
作者(Author): 王守强,朱大铭
参考文献(References):
- [1]Inaba M,Kaoth N,Imai H.Application of Weighted VoronoiDiagrams and Randomization to Variance-based k-cluste-ring:(Extended Abstract).In Proceeding of the Tenth An-nual Symposium on Computational Geometry,1994:332-339.
- [2]Amit Kumar,Yogish Saharwal,Sandeep Sen.A Sample Lin-ear Time(1+ε)Algorithm for k-Means Clustering in anyDimensions.2004:454-462.
- [3]Deerwester S C,Dumais S T,landauer T K,et al.Indexingby Latent Semantic Analysis[J].Amer J.Soc.Inform,1990,41(6):391-407.