Joint sparse principal component analysis
作者：
Highlights：
• The first contribution is JSPCA relaxes the orthogonal constraint to freely select the useful features.
• The second contribution is JSPCA integrates feature selection into subspace learning via joint l2,1-norms.
• The third contribution is JSPCA provides a simple yet effective optimization algorithm and a series of theoretical analyses.
摘要
Highlights•The first contribution is JSPCA relaxes the orthogonal constraint to freely select the useful features.•The second contribution is JSPCA integrates feature selection into subspace learning via joint l2,1-norms.•The third contribution is JSPCA provides a simple yet effective optimization algorithm and a series of theoretical analyses.
论文关键词：Dimensionality reduction,Joint sparse,ℓ2,1-norm
论文评审过程：Received 16 January 2016, Revised 21 August 2016, Accepted 22 August 2016, Available online 24 August 2016, Version of Record 2 September 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.08.025
