Nesting-structured nuclear norm minimization for spatially correlated matrix variate

作者：

Highlights：

• We take the local and global structures of a matrix variate into joint consideration.

• We propose a nesting-structured nuclear norm model and analyze its statistical meaning.

• We solve the proposed model by using an improved sub-gradient method.

• Experimental results demonstrate the advantages of our method.

摘要

•We take the local and global structures of a matrix variate into joint consideration.•We propose a nesting-structured nuclear norm model and analyze its statistical meaning.•We solve the proposed model by using an improved sub-gradient method.•Experimental results demonstrate the advantages of our method.

论文关键词：Nesting-structured nuclear norm,Low rank,Structured sparsity,Matrix regression,Matrix completion,Sub-gradient method

论文评审过程：Received 2 November 2016, Revised 30 January 2019, Accepted 18 February 2019, Available online 21 February 2019, Version of Record 26 February 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.02.011