Proximal Support Matrix Machine

In this paper, we have proposed a novel model called proximal support matrix machine (PSMM), which is mainly based on the models of vector (PSVM) and low rank (LRSMM). design, PSMM has comprehensively considered both relationship between samples same class structure rows or columns data. To certain extent, our can be regarded as synthesis PSVM LRSMM model. Since an unconstrained convex problem in essence, established alternating direction method multipliers algorithm to deal with Finally, since great experiments minst digital database show that classifier good ability distinguish two digits little difference, it encourages us conduct more complex MIT face database, INRIA person students Japan female facial expression database. Meanwhile, final experimental results performs better than PSVM, twin machine, linear multiple demanding image classification tasks.