Research on Support Vector Stepwise Regression Algorithm & its Improvement
支持向量逐步回归机及其改进算法研究

According to the nature that Support Vectors are sparse and they are located near a hyper-plane, Support Vectors Stepwise Regression Algorithm is proposed and its method of constructing a new sample subset was improved in this paper. The m is optimized with Integer Programming, which is number of new samples extracted in every search. Finally, analyzed its complexity, and tested and verified convergence find availability of the algorithms by the simulation.