Sparse matrix factorization with L2,1 norm for matrix completion
作者：
Highlights：
• We propose two matrix factorization methods DSMF and ISMF with l2,1 norm, where the former directly minimizes F2-norm loss function whiling the latter indirectly optimize the upper bound of F-norm function.
• We theoretically prove the convergence property of DSMF and discuss the convergence condition of ISMF.
• The experiments on on the simulation and benchmark datasets show that our methods achieve the comparable performance with the deep learning-based matrix completion methods.
摘要
•We propose two matrix factorization methods DSMF and ISMF with l2,1 norm, where the former directly minimizes F2-norm loss function whiling the latter indirectly optimize the upper bound of F-norm function.•We theoretically prove the convergence property of DSMF and discuss the convergence condition of ISMF.•The experiments on on the simulation and benchmark datasets show that our methods achieve the comparable performance with the deep learning-based matrix completion methods.
论文关键词：Matrix Completion,Matrix Factorization,L2,1 Norm Regularization,Alternative Optimization,Sparse Property
论文评审过程：Received 8 September 2020, Revised 3 March 2022, Accepted 14 March 2022, Available online 15 March 2022, Version of Record 21 March 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108655
