Robust Exponential Graph Regularization Non-Negative Matrix Factorization Technology for Feature Extraction

Graph regularized non-negative matrix factorization (GNMF) is widely used in feature extraction. In the process of dimensionality reduction, GNMF can retain internal manifold structure data by adding a regularizer to (NMF). Because Ga NMF implemented local preserving projections (LPP), there are small sample size problems (SSS). view above problems, new algorithm named robust exponential graph (REGNMF) proposed this paper. By exponent GNMF, possible existing singular will change into non-singular matrix. This model successfully solves algorithm. For optimization problem REGNMF algorithm, we use multiplicative updating rule iteratively solve method. Finally, method applied AR, COIL database, Yale noise set, and AR occlusion dataset for performance test, experimental results compared with some methods. The indicate that more significant.