支持向量分类器的模糊积分集成方法  

Support vector classifiers ensemble based on fuzzy integral

作　　者：颜根廷[1] 李传江[1] 马广富[1] 

机构地区：[1]哈尔滨工业大学航天学院

出　　处：《哈尔滨工业大学学报》2008年第7期1017-1020,共4页Journal of Harbin Institute of Technology

基　　金：教育部博士点专项基金资助项目(20050213010)

摘　　要：针对常规的基于投票方法的支持向量分类器影响集成分类器的泛化能力的问题,提出一种基于模糊积分的支持向量分类器集成方法,不仅考虑各子支持向量分类器输出的客观信息,同时还考虑各子分类器输出对于最终决策的重要性,提高了集成分类器的泛化能力.仿真试验表明,该方法的分类准确率明显优于单一支持向量分类器和传统基于投票方法的支持向量分类器集成策略.For the constraint of traditional voting-based support vector classifiers （SVCs） ensemble technique to the classification performance due to the impossibility of evaluating the importance degree of the output of individual component of SVC to the final decision, an SVCs ensemble method based on fuzzy integral is proposed. This method considers not only the objective information for the outputs of each component of SVC, but also the importance degree of the output of individual component of SVC to the final decision. Therefore, the classification performance is enhanced to a great extent. Simulation results demonstrate that the proposed SVCs ensemble approach based on fuzzy integral outperforms a single SVC and traditional SVCs ensemble technique via majority voting.

关 键 词：支持向量分类器 支持向量分类器集成 模糊积分 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]