Multi-view learning methods with the LINEX loss for pattern classification
作者：
Highlights：
• MVLSVM-CO and MVLSVM-SIM use LINEX loss to distinguish the error-prone sample.
• MVLSVM-CO and MVLSVM-SIM leverage the consistency and complementarity among views.
• An iterative two-step strategy is adopted to solve the optimization problems.
• The view-consistency and the generalization capability are theoretically analyzed.
• Comprehensive experiments verify the effectiveness of our proposed models.
摘要
•MVLSVM-CO and MVLSVM-SIM use LINEX loss to distinguish the error-prone sample.•MVLSVM-CO and MVLSVM-SIM leverage the consistency and complementarity among views.•An iterative two-step strategy is adopted to solve the optimization problems.•The view-consistency and the generalization capability are theoretically analyzed.•Comprehensive experiments verify the effectiveness of our proposed models.
论文关键词：Multi-view learning,Consensus and complementarity information,Asymmetric LINEX loss function,Support vector machine
论文评审过程：Received 23 January 2021, Revised 21 June 2021, Accepted 3 July 2021, Available online 6 July 2021, Version of Record 16 July 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107285
