Box-constrained projective nonnegative matrix factorization via augmented Lagrangian method.
Xiang ZhangNaiyang GuanLong LanDacheng TaoZhigang LuoPublished in: IJCNN (2014)
Keyphrases
- nonnegative matrix factorization
- augmented lagrangian method
- negative matrix factorization
- matrix factorization
- data representation
- total variation
- principal component analysis
- augmented lagrangian
- convex optimization
- constrained minimization
- least squares
- spectral clustering
- total variation regularization
- original data
- document clustering
- global convergence
- collaborative filtering
- data sets
- low rank
- sparse representation
- model selection
- objective function