Sketching for low-rank nonnegative matrix approximation: Numerical study

Abstract

We propose new approximate alternating projection methods, based on randomized sketching, for the low-rank nonnegative matrix approximation problem: find a low-rank approximation of a nonnegative matrix that is nonnegative, but whose factors can be arbitrary. We calculate the computational complexities of the proposed methods and evaluate their performance in numerical experiments. The comparison with the known deterministic alternating projection methods shows that the randomized approaches are faster and exhibit similar convergence properties.

arXiv commentAccepted version

References in corpus (5)

• Nonnegative Matrix Factorization for Signal and Data Analytics: Identifiability, Algorithms, and Applications
• Practical sketching algorithms for low-rank matrix approximation
• Efficient Nonnegative Tucker Decompositions: Algorithms and Uniqueness
• A Parallel Low-Rank Solver for the Six-Dimensional Vlasov-Maxwell Equations
• Low-rank nonnegative tensor approximation via alternating projections and sketching

Cited by in corpus (3)

• Nonnegative tensor train for the multicomponent Smoluchowski equation
• Quasioptimal alternating projections and their use in low-rank approximation of matrices and tensors
• When big data actually are low-rank, or entrywise approximation of certain function-generated matrices