Non-negative matrix factorization algorithm is a newly popular method for non-negative dimensionality reduction, feature extraction, data mining, etc. Further research on the algorithm of non-negative matrix factorization has its practical significance.Multiplicative update is a simple and popular way to the non-negative matrix factorization. In this paper, we propose a modified multiplicative update algorithm, in which we add some restricted items to the objective function. Based on optimization and some auxiliary function, we prove that the convergence of our new algorithm. From the eigenface obtained by this algorithm, one can observe that the local features of the eigenface will be stronger, as the parameter is larger. A image decomposed and reconstuceted by the algorithm has higher accuracy. |