主成分分析-支持向量机用于肝病分类模型  

Principal component analysis-support vector machine used to classified model of liver disease

作　　者：张云佳[1] 方亚平[1] 黄田海[2] 张婧[1] 谭福元[1] 孙李娜[1] 李梦龙[1] 

机构地区：[1]四川大学化学学院,四川成都610064 [2]解放军第三二四医院检验科,重庆400020

出　　处：《化学研究与应用》2008年第2期113-116,共4页Chemical Research and Application

基　　金：国家自然科学基金(29877016)资助项目

摘　　要：将主成分分析(PCA)用于肝功能检测数据特征提取,然后用支持向量机(SVM)对乙肝、丙肝、肝硬化、正常人样本建立分类模型。采用高斯径向基函数(RBF)为核函数,调节核函数参数C及σ以建立最佳支持向量机模型。该模型对训练集的识别率为99.3%,对预测集的预测率为96.4%。结果表明:PCA-SVM法建立的肝病分类模型能较好的区分乙肝、丙肝、肝硬化及正常人,且分类效果优于传统支持向量机及人工神经网络(ANN)分类模型。Principal component analysis（PCA） is used to feature selection of the liver function testing results, and the classified model of HBV,HCV,hepatocirrhosis patients and the normal is based on Support vector machine（SVM）. The radical basis function （RBF） is adopted as a kernel function of SVM,and the model adjusts C and σ to build the optimization classifier,which makes the correct classification ratio of the training set to be 99. 3%, while that of the testing set to be 96.4%. The result shows that the classified model of liver disease based on PCA-SVM can classifies the HBV, HCV, hepatocirrhosis patiems and the normal more effectively than the traditional SVM or ANN.

关 键 词：肝功能指标 主成分分析 支持向量机 人工神经网络 

分 类 号：O657[理学—分析化学]