改进的非线性RMPSVM及TMPSVM  

Improved Nonlinear RMPSVM and TMPSVM

作　　者：代永潇 李晓萌 范丽亚[1] DAI Yongxiao;LI Xiaomeng;FAN Liya(School of Mathematical Sciences,Liaocheng University,Liaocheng 252059,China)

机构地区：[1]聊城大学数学科学学院,山东聊城252059

出　　处：《聊城大学学报（自然科学版）》2022年第5期8-20,共13页Journal of Liaocheng University:Natural Science Edition

基　　金：国家自然科学基金项目(11801248);山东省自然科学基金项目(ZR2016AM24,ZR2018BF010)资助。

摘　　要：利用二次规划的Wolfe对偶形式改进了传统的非线性最小最大概率机(简记为非线性I-MPM),在此基础上借助于SVM思想提出了非线性正则化最小最大概率支持向量机(非线性RMPSVM)。为了弥补线性TMPMC和线性TMPM中没有考虑同类样本尽可能近的不足,结合非线性I-MPM和非线性TSVM的思想提出了非线性I-TMPSVM。实验结果表明非线性I-MPM和非线性RMPSVM的分类性能总体上优于非线性SVM,非线性I-TMPSVM不仅具有较低的计算复杂性,分类性能优于非线性TSVM和非线性I-MPM,且运行时间少于非线性TSVM。In this paper,the traditional nonlinear minimax probability machine(nonlinear I-MPM)is improved by using the Wolfe duality form of quadratic programming.On this basis,a nonlinear regularized minimax probability support vector machine(nonlinear RMPSVM)is proposed by using the idea of SVM.In order to make up for the deficiency that linear TMPMC and linear TMPM do not consider similar samples as close as possible,this paper proposes nonlinear I-TMPSVM combining nonlinear I-MPM and nonlinear TSVM.The experimental results show that the classification performance of nonlinear I-MPM and nonlinear RMPSVM is generally better than that of nonlinear SVM.The nonlinear I-TMPSVM not only has lower computational complexity,but also has better classification performance than that of nonlinear TSVM and nonlinear I-MPM,and the running time is less than that of nonlinear TSVM.

关 键 词：最小最大概率机 支持向量机 孪生支持向量机 孪生最小最大概率支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]