Abstract: Non-negative matrix factorisation (NMF or NNMF) is a group of algorithms for matrix decomposition with the non-negative constraint, that is the matrix to be factorised and the factors of the matrix should be non-negative. NMF is a very popular technique for data decomposition. Some of the data’s such as image pixels, power spectra, stock indices are non- negative. NMF has found wide variety of applications mostly in pre-processing of signals and classification. So in this paper we review about the concepts of NMF and its applications in various fields.

Keywords: NMF, signal processing, power spectra, classification.