Abstract: Nonncgativc matrix factorization(NMI)is based on part feature extraction algorithm which adds nonnegative constraint into matrix factorization. A method called graph regularized non-negative matrix factorization with sparseness constraints(UNMFSC)was proposed for enhancing the classification accuracy. It not only considers the geometric structure in the data representation, but also introduces sparseness constraint to coefficient matrix and integrates them into one single objective function. An efficient multiplicative updating procedure was produced along with its theoretic justificanon of the algorithmic convergence. Experiments on ORI. and MI T-CBCI. face recognition databases demonstrate the effectiveness of the proposed method.

Key words: Non-negative matrix, Graph Regularization, Sparse coding