Auto-weighted multi-view constrained spectral clustering

Auto-weighted multi-view constrained spectral clustering
复制标题

自动加权多视图约束谱聚类

DOI:
10.1016/j.neucom.2019.06.098
复制
发表时间:
2019-11
期刊:
影响因子:
6
通讯作者:
Zhu Hong
Zhu Hong
中科院分区:
计算机科学2区
文献类型:
--
作者:
Chen Chuan;Qian Hui;Chen Wuhui;Zheng Zibin;Zhu Hong

文献摘要

参考文献

被引文献

相似文献

约束聚类是一种新的半监督学习方式,它通过利用成对约束来提高划分的质量。虽然许多约束聚类方法在单视图聚类中具有很好的性能,但它们不能直接应用于多视图场景。在本文中,我们提出了一种新的约束谱聚类方法的多视图数据,明确规定成对的约束作为一系列的线性约束的统一指标矩阵。据我们所知,这是第一个多视图约束谱聚类的工作。我们的方法可以通过自动权重学习策略来区分不同视图的重要性。同时,包含大量噪声或无关信息的视图也被自动消除,从而提高预测性能。在不同多视图数据集上进行的大量实验表明,该方法可以有效地利用成对约束,并优于最先进的方法。
Constrained clustering is a new fashion of semi-supervised learning which focused on enhancing the quality of the partition by utilizing pairwise constraints. Though many constrained clustering methods have an excellent performance in single-view clustering, they can’t be directly applied to multi-view scenario. In this paper, we propose a novel constrained spectral clustering approach for multi-view data, which explicitly imposes pairwise constraints as a series of linear constraints on the unified indicator matrix. To our best knowledge, this is the first work on multi-view constrained spectral clustering. Our approach can differ the importance of different views via the auto-weight learning strategy. Simultaneously, the views which contain much noisy or irrelevant information are also automatically eliminated, thereby improving the prediction performance. Extensive experiments conducted on various multi-view datasets demonstrate that the proposed approach can efficiently utilize pairwise constraints and outperforms the state-of-the-art approaches.
DOI: 10.1145/1835804.1835877
发表时间: 2010-07
期刊: Proceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and data mining
影响因子: --
作者:
Xiang Wang;I. Davidson
通讯作者: Xiang Wang;I. Davidson
DOI: 10.1109/icassp.2008.4518374
发表时间: 2008-05
影响因子: 5.4
作者:
Stephen J. Wright;R. Nowak;Mário A. T. Figueiredo
通讯作者: Stephen J. Wright;R. Nowak;Mário A. T. Figueiredo
DOI: 10.5445/ir/1000011477
发表时间: 2007
期刊: --
影响因子: --
作者:
Silke Wagner;D. Wagner
通讯作者: Silke Wagner;D. Wagner
DOI: 10.1007/978-3-642-15567-3_1
发表时间: 2010-09
期刊: --
影响因子: --
作者:
Zhiwu Lu;H. Ip
通讯作者: Zhiwu Lu;H. Ip
DOI: 10.5555/1756006.1756037
发表时间: 2010-03
期刊: J. Mach. Learn. Res.
影响因子: --
作者:
Andreas Argyriou;C. Micchelli;M. Pontil
通讯作者: Andreas Argyriou;C. Micchelli;M. Pontil