一种多类支持向量机概率建模新方法

New Method About Probability Modeling of Multi-class Support Vector Machines

在支持向量机多类分类问题输出概率建模中,提出了一种直接求解后验概率的概率建模新方法。在对多个两类支持向量机分类器的输出概率进行组合时,该方法充分考虑了各个两类支持向量机分类器的差异,并以后验概率作为各个两类支持向量机分类器的权系数。仿真图像的实验结果表明,该文提出的直接求解后验概率方法与投票法及PairwiseCoupling方法相比,不仅具有较好的分类性能,而且得到的后验概率具有较好的概率分布形态。

A directly solving posterior probability method is presented for probability output of SVM in the multi-class case. The differences and different weights among these two-class SVM classifiers, based on the posterior probability, are considered and given for the combination of the probability outputs. The simulation experiment results show that the directly solving posterior probability method achieves the better classification ability and the better probability distribution of the posterior probability than the voting method and the Pairwise Coupling method.