Simplicial nonnegative matrix factorization

Nonnegative matrix factorization (NMF) plays a crucial role in machine learning and data mining, especially for dimension reduction and component analysis. It is employed widely in different fields such as information retrieval, image processing, etc. After a decade of fast development, severe limitations still remained in NMFs methods including high complexity in instance inference, hard to control sparsity or to interpret the role of latent components. To deal with these limitations, this paper proposes a new formulation by adding simplicial constraints for NMF. Experimental results in comparison to other state-of-the-art approaches are highly competitive.