By using an existence theorems of maximal elements for a family of set-valued mappings in G-convex spaces due to the author, some new nonempty intersection theorems for a family of set-valued mappings were established...By using an existence theorems of maximal elements for a family of set-valued mappings in G-convex spaces due to the author, some new nonempty intersection theorems for a family of set-valued mappings were established in noncompact product G-convex spaces. As applications, some equilibrium existence theorems for a system of generalized vector equilibrium problems were proved in noncompact product G-convex spaces. These theorems unify, improve and generalize some important known results in literature.展开更多
对支持向量机(Twin Support Vector Machine,TWSVM)的优化思想源于基于广义特征值近似支持向量机(ProximalSVM based on Generalized Eigenvalues,GEPSVM)。该算法将传统SVM问题分解为两个凸规划问题,使得训练速度缩减到原来的1/4。对TW...对支持向量机(Twin Support Vector Machine,TWSVM)的优化思想源于基于广义特征值近似支持向量机(ProximalSVM based on Generalized Eigenvalues,GEPSVM)。该算法将传统SVM问题分解为两个凸规划问题,使得训练速度缩减到原来的1/4。对TWSVM做了修正,基于新的优化准则设计了一种特殊TWSVM(GTWSVM),在此基础上,提出了快速GTWSVM(FGTWSVM),其将GTWSVM转换为无约束凸规划问题求解。该算法在保证得到与TWSVM相当的分类性能以及较快的计算速度的同时,还减少了输入空间的特征数以及内存占用。对于非线性问题,FGTWSVM可以减少核函数数目。展开更多
文摘By using an existence theorems of maximal elements for a family of set-valued mappings in G-convex spaces due to the author, some new nonempty intersection theorems for a family of set-valued mappings were established in noncompact product G-convex spaces. As applications, some equilibrium existence theorems for a system of generalized vector equilibrium problems were proved in noncompact product G-convex spaces. These theorems unify, improve and generalize some important known results in literature.
基金Supported by the Guizhou Province Natural Science Foundation of China([2011]2093)the Natural Scientific Research Foundation of Guizhou Provincial Education Department((2012)058)
文摘对支持向量机(Twin Support Vector Machine,TWSVM)的优化思想源于基于广义特征值近似支持向量机(ProximalSVM based on Generalized Eigenvalues,GEPSVM)。该算法将传统SVM问题分解为两个凸规划问题,使得训练速度缩减到原来的1/4。对TWSVM做了修正,基于新的优化准则设计了一种特殊TWSVM(GTWSVM),在此基础上,提出了快速GTWSVM(FGTWSVM),其将GTWSVM转换为无约束凸规划问题求解。该算法在保证得到与TWSVM相当的分类性能以及较快的计算速度的同时,还减少了输入空间的特征数以及内存占用。对于非线性问题,FGTWSVM可以减少核函数数目。
