正在加载图片...
.424. 智能系统学报 第11卷 100 sing evidence accumulation.IEEE transactions on pattern 一米一本文方法 analysis and machine intelligence,2005,27(6):835-850. -eEAC 80 X一HBGF -SRS 0 [5]WANG Xi,YANG Chunyu,ZHOU Jie.Clustering aggrega- 米 -·米-WCT tion by probability accumulation[J].Pattern recognition, -米-WEAC 60 -·e-GP-MGLA 2009,42(5):668-675 [6]SINGH V,MUKHERJEE L,PENG Jiming,et al.Ensemble % -8.0. 米 clustering using semidefinite programming with applications [J].Machine learning,2010.79(1/2):177-200. [7]HUANG Dong,LAI Jianhuang,WANG Changdong.Exploi- 杀举 米 ×109 ting the wisdom of crowd:a multi-granularity approach to 0.5 1.0 1.5 2.0 clustering ensemble[C]//Proceedings of the 4th Internation- 数据规模 al Conference on Intelligence Science and Big Data Engineer- (h)LR ing.Beijing,China,2013:112-119. 图4各个聚类集成方法在不同数据规模下的运行时间 [8]HUANG Dong,LAI Jianhuang,WANG Changdong.Combi- 对比 ning multiple clusterings via crowd agreement estimation and Fig.4 Execution time of different methods with varying multi-granularity link analysis[J].Neurocomputing,2015, data sizes 170:240-250. 3 结束语 [9]HUANG Dong,LAI Jianhuang,WANG Changdong.Ensem- 为解决聚类集成研究中的聚类成员可靠度估计 ble clustering using factor graph[J].Pattern recognition, 2016,50:131-142 与加权问题,本文提出了一个基于二部图结构与决策 [10]HUANG Dong,LAI Jianhuang,WANG Changdong.Robust 加权机制的聚类集成方法。我们将每个聚类成员视 ensemble clustering using probability trajectories[J].IEEE 作一个包含若干连接决策的集合,并为每个聚类成员 transactions on knowledge and data engineering,2016,28 的决策集合分配一个单位的可信度。该可信度由聚 (5):1312-1326. 类成员内的各个决策共同分享。进一步地,我们提出 [11]LI Tao,DING C.Weighted consensus clustering[C]//Pro- 基于可信度分享的决策加权机制,并将之整合至一个 ceedings of the 2008 SIAM International Conference on Data 统一的二部图模型中。因其二部图结构,该图模型可 mining.Auckland,New Zealand,2008:798-809. 利用Tcut算法进行快速分割,从而得到最终聚类集[l2]KARYPIS G,KUMAR V.Multilevel k-way partitioning 成结果。本文在8个实际数据集中进行了实验,将所 scheme for irregular graphs[J].Journal of parallel and dis- 提出方法与聚类成员以及6个现有方法进行了对比 tributed computing,1998,48(1):96-129. 分析。实验结果验证了本文方法在聚类质量及运算 [13]NG A Y,JORDAN M I,WEISS Y.On spectral clustering: Analysis and an algorithm[C]//Advances in Neural Infor- 效率上的显著优势。 mation Processing Systems.Vancouver,Canada,2001. 参考文献: [14]TOPCHY A,JAIN A K,PUNCH W.Clustering ensembles: models of consensus and weak partitions[J.IEEE transac- [1]STREHL A,GHOSH J.Cluster ensembles-a knowledge reuse tions on pattern analysis and machine intelligence,2005,27 framework for combining multiple partitions[J].The journal (12):1866-1881. of machine learning research,2003,3(3):583-617. [15]VEGA-PONS S,CORREA-MORRIS J,RUIZ-SHULCLOP- 2]CRISTOFOR D,SIMOVICI D.Finding median partitions u- ER J.Weighted partition consensus via kernels[]].Pattern sing information-theoretical-based genetic algorithms [J]. rec0 gnition,2010,43(8):2712-2724. Journal of universal computer science,2002,8(2):153-[16 VEGA-PONS S,RUIZ-SHULCLOPER J,GUERRA- 172 GANDON A.Weighted association based methods for the [3]FERN X Z,BRODLEY C E.Solving cluster ensemble prob- combination of heterogeneous partitions[J].Pattern recog- lems by bipartite graph partitioning[C]//Proceedings of the nition letters,.2011,32(16):2163-2170, 2 Ist International Conference on Machine Learning.New[I7]徐森,周天,于化龙,等.一种基于矩阵低秩近似的聚类 York,NY,USA,2004. 集成算法[J].电子学报,2013,41(6):1219-1224. [4]FRED A L N,JAIN A K.Combining multiple clusterings u- XU Sen,ZHOU Tian,YU Hualong,et al.Matrix low rank(h)LR 图 4 各个聚类集成方法在不同数据规模下的运行时间 对比 Fig.4 Execution time of different methods with varying data sizes 3 结束语 为解决聚类集成研究中的聚类成员可靠度估计 与加权问题,本文提出了一个基于二部图结构与决策 加权机制的聚类集成方法。 我们将每个聚类成员视 作一个包含若干连接决策的集合,并为每个聚类成员 的决策集合分配一个单位的可信度。 该可信度由聚 类成员内的各个决策共同分享。 进一步地,我们提出 基于可信度分享的决策加权机制,并将之整合至一个 统一的二部图模型中。 因其二部图结构,该图模型可 利用 Tcut 算法进行快速分割,从而得到最终聚类集 成结果。 本文在 8 个实际数据集中进行了实验,将所 提出方法与聚类成员以及 6 个现有方法进行了对比 分析。 实验结果验证了本文方法在聚类质量及运算 效率上的显著优势。 参考文献: [1]STREHL A, GHOSH J. Cluster ensembles⁃a knowledge reuse framework for combining multiple partitions[ J]. The journal of machine learning research, 2003, 3(3): 583⁃617. [2]CRISTOFOR D, SIMOVICI D. Finding median partitions u⁃ sing information⁃theoretical⁃based genetic algorithms [ J ]. Journal of universal computer science, 2002, 8 ( 2): 153⁃ 172. [3]FERN X Z, BRODLEY C E. Solving cluster ensemble prob⁃ lems by bipartite graph partitioning[C] / / Proceedings of the 21st International Conference on Machine Learning. New York, NY, USA, 2004. [4]FRED A L N, JAIN A K. Combining multiple clusterings u⁃ sing evidence accumulation[J]. IEEE transactions on pattern analysis and machine intelligence, 2005, 27(6): 835⁃850. [5]WANG Xi, YANG Chunyu, ZHOU Jie. Clustering aggrega⁃ tion by probability accumulation [ J ]. Pattern recognition, 2009, 42(5): 668⁃675. [6]SINGH V, MUKHERJEE L, PENG Jiming, et al. Ensemble clustering using semidefinite programming with applications [J]. Machine learning, 2010, 79(1 / 2): 177⁃200. [7] HUANG Dong, LAI Jianhuang, WANG Changdong. Exploi⁃ ting the wisdom of crowd: a multi⁃granularity approach to clustering ensemble[C] / / Proceedings of the 4th Internation⁃ al Conference on Intelligence Science and Big Data Engineer⁃ ing. Beijing, China, 2013: 112⁃119. [8] HUANG Dong, LAI Jianhuang, WANG Changdong. Combi⁃ ning multiple clusterings via crowd agreement estimation and multi⁃granularity link analysis [ J]. Neurocomputing, 2015, 170: 240⁃250. [9] HUANG Dong, LAI Jianhuang, WANG Changdong. Ensem⁃ ble clustering using factor graph [ J ]. Pattern recognition, 2016, 50: 131⁃142. [10]HUANG Dong, LAI Jianhuang, WANG Changdong. Robust ensemble clustering using probability trajectories[ J]. IEEE transactions on knowledge and data engineering, 2016, 28 (5): 1312⁃1326. [11]LI Tao, DING C. Weighted consensus clustering[C] / / Pro⁃ ceedings of the 2008 SIAM International Conference on Data mining. Auckland, New Zealand, 2008: 798⁃809. [ 12 ] KARYPIS G, KUMAR V. Multilevel k⁃way partitioning scheme for irregular graphs[J]. Journal of parallel and dis⁃ tributed computing, 1998, 48(1): 96⁃129. [13]NG A Y, JORDAN M I, WEISS Y. On spectral clustering: Analysis and an algorithm[C] / / Advances in Neural Infor⁃ mation Processing Systems. Vancouver, Canada, 2001. [14]TOPCHY A, JAIN A K, PUNCH W. Clustering ensembles: models of consensus and weak partitions[ J]. IEEE transac⁃ tions on pattern analysis and machine intelligence, 2005, 27 (12): 1866⁃1881. [15]VEGA⁃PONS S, CORREA⁃MORRIS J, RUIZ⁃SHULCLOP⁃ ER J. Weighted partition consensus via kernels[ J]. Pattern recognition, 2010, 43(8): 2712⁃2724. [ 16 ] VEGA⁃PONS S, RUIZ⁃SHULCLOPER J, GUERRA⁃ GANDóN A. Weighted association based methods for the combination of heterogeneous partitions[J]. Pattern recog⁃ nition letters, 2011, 32(16): 2163⁃2170. [17]徐森, 周天, 于化龙, 等. 一种基于矩阵低秩近似的聚类 集成算法[J]. 电子学报, 2013, 41(6): 1219⁃1224. XU Sen, ZHOU Tian, YU Hualong, et al. Matrix low rank ·424· 智 能 系 统 学 报 第 11 卷
<<向上翻页向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有