Multi-class feature selection by exploring reliable class correlation
作者：
Highlights：
• Preserving the correlation of different classes in the process of feature selection.
• Gradually inducing reduction feature space to approximate target feature space to capture reliable class correlation.
• Simultaneously learning similarity matrix P and a feature weighting matrix W to avoid the influence of redundant and irrelevant features as much as possible.
• Comprehensive experiments on both synthetic data and various public benchmark datasets.
摘要
•Preserving the correlation of different classes in the process of feature selection.•Gradually inducing reduction feature space to approximate target feature space to capture reliable class correlation.•Simultaneously learning similarity matrix P and a feature weighting matrix W to avoid the influence of redundant and irrelevant features as much as possible.•Comprehensive experiments on both synthetic data and various public benchmark datasets.
论文关键词：Multi-class learning,Feature selection,Reliable class correlation,Dimension reduction feature space
论文评审过程：Received 29 March 2021, Revised 5 August 2021, Accepted 10 August 2021, Available online 12 August 2021, Version of Record 26 August 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107377
