Adaptive weighted nonnegative low-rank representation
作者：
Highlights：
• In this paper, a novel graph learning method is proposed to learn a more interpretable and robust graph for data clustering.
• The global and local structures are simultaneously exploited for graph learning, which ensures to learn a more reasonable graph.
• The learned graph has good interpretability to samples by introducing the nonnegative constraint to the model.
• The proposed method is robust to noise and redundant features by introducing an adaptive weighted matrix.
摘要
•In this paper, a novel graph learning method is proposed to learn a more interpretable and robust graph for data clustering.•The global and local structures are simultaneously exploited for graph learning, which ensures to learn a more reasonable graph.•The learned graph has good interpretability to samples by introducing the nonnegative constraint to the model.•The proposed method is robust to noise and redundant features by introducing an adaptive weighted matrix.
论文关键词：Low-rank representation,Adaptive weighted matrix,Data clustering,Locality constraint
论文评审过程：Received 26 May 2017, Revised 12 December 2017, Accepted 4 April 2018, Available online 11 April 2018, Version of Record 18 April 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.04.004
