L1-norm and maximum margin criterion based discriminant locality preserving projections via trace Lasso
作者：
Highlights：
• We propose a L1-norm and MMC based DLPP method via trace Lasso.
• Our proposed algorithm can simultaneously consider the sparsity and correlation.
• We also propose an efficient procedure for solving the proposed method.
摘要
•We propose a L1-norm and MMC based DLPP method via trace Lasso.•Our proposed algorithm can simultaneously consider the sparsity and correlation.•We also propose an efficient procedure for solving the proposed method.
论文关键词：Discriminant locality preserving projections,Feature extraction,Maximum margin criterion,L1-norm,Trace Lasso,L2-norm
论文评审过程：Received 9 April 2015, Revised 18 January 2016, Accepted 28 January 2016, Available online 4 February 2016, Version of Record 21 March 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.01.029
