Lower Bounds on Matrix Factorization Ranks via Noncommutative Polynomial Optimization.
Sander GriblingDavid de LaatMonique LaurentPublished in: Found. Comput. Math. (2019)
Keyphrases
- matrix factorization
- lower bound
- collaborative filtering
- stochastic gradient descent
- low rank
- recommender systems
- factor analysis
- missing data
- nonnegative matrix factorization
- factorization methods
- negative matrix factorization
- data sparsity
- variational bayesian
- multiscale
- implicit feedback
- constrained optimization
- ranking functions
- personalized ranking
- higher order
- denoising
- factorization method
- tensor factorization
- low rank matrix
- objective function
- learning algorithm
- item recommendation
- factor matrices
- probabilistic matrix factorization