Structure regularized sparse coding for data representation

作者：

Highlights：

• We discuss the drawbacks of the local affinity matrix and build the global matrix.

• We propose a novel unsupervised method for the sparse representation of the data.

• The proposed method exploits the latent discriminant information in the data

• We develop an ADMM-based iterative optimization method to solve the proposed model.

摘要

•We discuss the drawbacks of the local affinity matrix and build the global matrix.•We propose a novel unsupervised method for the sparse representation of the data.•The proposed method exploits the latent discriminant information in the data•We develop an ADMM-based iterative optimization method to solve the proposed model.

论文关键词：Data representation,Sparse coding,Graph regularized,Unsupervised learning

论文评审过程：Received 23 September 2018, Revised 24 February 2019, Accepted 28 February 2019, Available online 6 March 2019, Version of Record 18 April 2019.

论文官网地址：https://doi.org/10.1016/j.knosys.2019.02.035