Self-paced non-convex regularized analysis–synthesis dictionary learning for unsupervised feature selection

作者：

Highlights：

• A novel UFS approach is proposed by incorporating self-paced and non-convex sparse regularization into dictionary learning.

• An effective iterative algorithm based on alternative search strategy is developed.

• The theoretical properties of the optimization on the convergence and computational complexity are investigated.

• Extensive experiments are conducted to verify the effectiveness and superiority of the proposed method.

摘要

•A novel UFS approach is proposed by incorporating self-paced and non-convex sparse regularization into dictionary learning.•An effective iterative algorithm based on alternative search strategy is developed.•The theoretical properties of the optimization on the convergence and computational complexity are investigated.•Extensive experiments are conducted to verify the effectiveness and superiority of the proposed method.

论文关键词：Unsupervised feature selection,Self-paced learning,Non-convex regularization,Alternative search strategy

论文评审过程：Received 25 November 2021, Revised 12 January 2022, Accepted 21 January 2022, Available online 31 January 2022, Version of Record 10 February 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108279