摘要
针对如何利用多视图数据中的隐含信息以及避免后续处理过程中带来的聚类性能次优问题,提出一种基于一致性和差异性的低秩张量多视图聚类算法.首先,该算法同时考虑视图的一致性和差异性信息,将多个一致性相似矩阵叠加在一个受低秩约束的张量中,以探索视图间信息的高阶相关性,从而得到更高质量的相似矩阵;其次,通过学习一个一致非负嵌入矩阵直接获得聚类结果;再次,采用自适应加权策略考虑不同视图数据的重要性;最后,通过在6个真实数据集上与其他算法进行对比实验,验证了该算法在多视图聚类问题上的有效性.
Aiming at how to utilize the implicit information in multi-view data and avoid the problem of sub-optimal clustering performance in the subsequent processing,we proposed a multi-view clustering algorithm with low-rank tensor based on consistency and difference.Firstly,the algorithm simultaneously considered the consistency and differential information of views,and superimposed multiple consistent similarity matrices in a tensor constrainted by low-rank to explore the higher-order correlations of the information between views,thus obtaining higher-quality similarity matrices.Secondly,clustering results were directly obtained by learning a consistent non-negative embedding matrix.Thirdly,an adaptive weighting strategy was used to consider the importance of different view data.Finally,the effectiveness of the algorithm on the multi-view clustering problem was verified by comparison experiments with other algorithms on six real datasets.
作者
周余琳
王长鹏
ZHOU Yulin;WANG Changpeng(School of Sciences,Chang’an University,Xi’an 710064,China)
出处
《吉林大学学报(理学版)》
北大核心
2025年第2期537-550,共14页
Journal of Jilin University:Science Edition
基金
国家自然科学基金(批准号:12471480)
长安大学中央高校基本科研业务费专项基金(批准号:300102122101)。
关键词
多视图聚类
一致性
差异性
低秩张量表示
自适应加权
multi-view clustering
consistency
difference
low-rank tensor representation
adaptive weighting
