摘要
核化一类硬划分SVDD、一/二类L2-SVM、L2支持向量回归和Ranking SVM均已被证明是中心约束最小包含球.这里将多视角学习引入核化L2-SVM,提出核化两类多视角L2-SVM(Multi-view L2-SVM),并证明该核化两类Multi-view L2-SVM亦为中心约束最小包含球,进而提出一种多视角核心向量机Mv CVM.所提出的Multi-view L2-SVM和Mv CVM既考虑了视角之间的差异性,又考虑了视角之间的关联性,使得分类器在各个视角上的学习结果趋于一致.人造多视角数据集和真实多视角数据集的实验均表明了Multi-view L2-SVM和Mv CVM方法的有效性.
The kernelized one-class hard-margin SVDD, the kernelized soft-margin one-class and two-class SVMs, the kernelized L2-support vector regression, and the kernelized Ranking SVM can be proved to be the center-constrained minimum enclosing ball(CCMEB) problem. Therefore, a kernelized two-class L2-SVM with multi-view(multi-view L2-SVM) is equivalently formulated as the CCMEB problem, and a classification method named multi-view core vector machine(Mv CVM) is proposed. Both the proposed multi-view L2-SVM and Mv CVM classifiers can obtain an overall consensus classification result on each view because the differences and the associations between different views are both considered in the two proposed models. An extensive set of experiments on synthetic and real-world multi-view datasets are conducted to demonstrate the effectiveness of the proposed methods.
出处
《控制与决策》
EI
CSCD
北大核心
2015年第8期1356-1364,共9页
Control and Decision
基金
国家自然科学基金项目(61272210
61202311)
江苏省自然科学基金项目(BK2012552)
贵州省科学技术基金项目(黔科合J字[2013]2136号
黔科合J字LKM[2013]23)
关键词
多视角
视角差异性
视角关联性
一致性
核心向量机
multi-view
differences of different views
associations of different views
consensus
core vector machine