A density-based similarity matrix construction for spectral clustering

被引:29
作者
Beauchemin, Mario [1 ]
机构
[1] Nat Resources Canada, Canada Ctr Remote Sensing, Ottawa, ON K1A 0E4, Canada
关键词
Affinity matrix; Non-parametric density estimation; Spectral clustering; K-means algorithm; Cluster ensembles; AGGREGATION;
D O I
10.1016/j.neucom.2014.10.012
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In the first part of this paper, we present a method to build affinity matrices for spectral clustering from a density estimator relying on K-means with subbagging procedure. The approach is anchored in the theoretical works of Wong (1980, 1982a, b) [13-15] on the asymptotic properties of K-means as a density estimation method. The subbagging procedure is introduced to improve the density estimate accuracy. The behavior of the proposed method is analyzed on diverse data sets and two new mechanisms are suggested to improve clustering results on non-convex data. In the second part of the paper, we establish a link between the presented method and the evidence accumulation clustering (EAC) approach by showing that a normalized version of the density-based similarity matrix is approximately equal to a normalized version of the co-association matrix. The co-association matrix provides the co-occurrence probability of data pairs assigned to a same cluster over multiple K-means clustering partitions. Experimental results on artificial and real data demonstrate the effectiveness of the method and provide empirical support for the established link. Crown Copyright (C) 2014 Published by Elsevier B.V. All rights reserved.
引用
收藏
页码:835 / 844
页数:10
相关论文
共 35 条
[1]  
Aidos H, 2011, LECT NOTES COMPUT SC, V7005, P290, DOI 10.1007/978-3-642-24471-1_21
[2]  
Andonova S, 2002, FR ART INT, V77, P513
[3]  
[Anonymous], P 20 INT C INF INT I
[4]  
[Anonymous], COMPUT SCI STAT
[5]  
[Anonymous], COMPUT SCI STAT
[6]  
[Anonymous], P 31 IEEE INT C AC S
[7]  
[Anonymous], KENDALLS ADV THEORY
[8]  
[Anonymous], 200080 MIT SLOAN SCH
[9]  
[Anonymous], 134082 MIT SLOAN SCH
[10]  
[Anonymous], THESIS G MASON U