Weighted Nuclear Norm Minimization Model for Matrix Completion
Collaborative filtering is one of the popular techniques used in recommendation system.It has some advantages over traditional recommendation technologies.But the limitation is that it is constrained by the data sparsity.Matrix completion technology can be used to solve this problem.This paper proposed an weighted nuclear norm minimization (WNNM) model for matrix completion to improve the flexibility of nuclear norm.Under certain condition,it can be proved to get global optimal solution.Meanwhile,convergence for another form of the proposed model was confirmed.With two real data sets,convex optimization algorithm of nuclear norm minimization was achieved to verify the proposed model.The result proves that to some extent it improves the computational speed and accuracy.
NETL
NSTL
万方数据
维普
