A two-point step size gradient method fornon-negative matrix factorization

Nonnegative Matrix Factorization (NMF) is a recently developed technique for finding parts-based, linear representations of non-negative data. There are many methods for NMF, such as multiplicative iterative and alternating nonnegative data least squares algorithms (ANLS). Among these methods, the most effective way is alternating nonnegative data least squares algorithms. In this paper, we propose a two-point step size gradient method to solve sub-problems of ANLS. The numerical experiment shows that our algorithm is much more efficient than some existing algorithms.