一种可能性测度的模糊支持向量机  

Fuzzy support vector machine based on possibility measure

作　　者：沈菊红[1] 黄永东[1] 

机构地区：[1]北方民族大学信息与计算科学学院,银川750021

出　　处：《黑龙江大学自然科学学报》2012年第2期204-208,共5页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(10961001);宁夏自然科学基金资助项目(NZ1056);北方民族大学科学研究基金资助项目(2009Y032)

摘　　要：支持向量机能够成功的解决分类和回归问题,但是训练数据都是精确的。如果支持向量机的训练集中含有模糊信息,即训练集中的输入训练样本点为模糊数,那么支持向量机将无能为力。基于此,在可能性测度理论和模糊机会约束规划的基础上,建立了模糊v-支持向量机模型,并将该方法应用于某病的诊断中,实验结果验证了该方法的有效性。Support vector machines have been very successful in classification and regression problems,but the training data are non-fuzzy.There are lots of fuzzy information in the objective world,and if there are fuzzy information in the training set of support vector machines（SVMs）,that is the input of training point is fuzzy number,then traditional SVMs will fail.Fuzzy v-support vector machine based on possibility measure and fuzzy chance constrained programming is proposed.Finally,fuzzy v-support vector machine is applied to the coronary ill diagnose,and the experimental result shows that the proposed model is promising.

关 键 词：支持向量机 模糊数 可能性测度 模糊机会约束规划 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]