2019/03/04 by Paola Favati, Favati, Paola, G. Lotti +5
Computer Science · #Face and Expression Recognition
paper · pdf · doi:10.48550/arxiv.1903.01321
Nonnegative Matrix Factorization (NMF), first proposed in 1994 for data\nanalysis, has received successively much attention in a great variety of\ncontexts such as data mining, text clustering, computer vision, bioinformatics,\netc. In this paper the case of a symmetric matrix is considered and the\nsymmetric nonnegative matrix factorization (SymNMF) is obtained by using a\npenalized nonsymmetric minimization problem. Instead of letting the penalizing\nparameter increase according to an a priori fixed rule, as suggested in\nliterature, we propose a heuristic approach based on an adaptive technique.\nExtensive experimentation shows that the proposed algorithm is effective.\n