Alternating proximal gradient method for nonnegative matrix factorization

Details Alternating proximal gradient method for nonnegative matrix factorization Alternating proximal gradient method for nonnegative matrix factorization

2011-12-22

arxiv.org/abs/1112.5407v1

Computer Science

Information Theory

Scientific paper

Nonnegative matrix factorization has been widely applied in face recognition, text mining, as well as spectral analysis. This paper proposes an alternating proximal gradient method for solving this problem. With a uniformly positive lower bound assumption on the iterates, any limit point can be proved to satisfy the first-order optimality conditions. A Nesterov-type extrapolation technique is then applied to accelerate the algorithm. Though this technique is at first used for convex program, it turns out to work very well for the non-convex nonnegative matrix factorization problem. Extensive numerical experiments illustrate the efficiency of the alternating proximal gradient method and the accleration technique. Especially for real data tests, the accelerated method reveals high superiority to state-of-the-art algorithms in speed with comparable solution qualities.

Affiliated with

Also associated with

No associations

Rating

Alternating proximal gradient method for nonnegative matrix factorization does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.

If you have personal experience with Alternating proximal gradient method for nonnegative matrix factorization, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Alternating proximal gradient method for nonnegative matrix factorization will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFWR-SCP-O-308203