基于决策树的代偿期和失代偿期肝硬化自动诊断的方法.pdfVIP

：!堑： 现岱生物医堂进展婴!堕!坠垫坐壁亟!塾壁±!￡骢里理窿壁垦堕垫M垒巫曼堡坠地堕壁亟鲤鱼壁垒Q竖盟丛!曼盥Q!! 基于决策树的代偿期和失代偿期肝硬化自动诊断的方法术 惠2 王彦1△马利庄1’2褚娜1周 摘要：决策树方法因结构简单、便于理解和具有较高的分类精度而在数据挖掘中被广泛采用。本文利用改进的决策树算法c4．5从 201例肝硬化病例中自动地提取相应的肝硬化状态识别规则，得到决策树分类模型并归纳出代偿性肝硬化和失代偿性肝硬化的 诊断规则，识别正确率为84．6％。实验结果表明决策树能较好的自动从肝硬化病例中归纳出代偿性和失代偿性肝硬化的诊断规 则。 关键词：决策树；肝硬化；数据挖掘；代偿期；失代偿期 中图分类号：R575．2文献标识码：A文章编号：1673—6273(2008)01一0126-03 MethodonDecisionfor Based Tree Diagnostic Compensated and LiverCimlosis米 Decompensated wANG M证 M．zHoU Ya矗凸，MALi—zhuaIl酽，CHU (1 D印a舳entofCo嘲putefScicnceEn舀ne砸ngShan曲aiJiaotonguniversi彤Shan曲ai200240，Ch抽a： 2Cen把rof ABSllRACT：Decision-treecalculationhasbeen usedinthedata becauseofits stnlctureand widely mining simple easy aswellas classification meansof decision一仃eecalculation ing higher precision．In也is锄icle，byimpmved C4．5，correspondingrecogni· tion 201 ci玎hosiswere ex仃acted舶mcases modelsof pmciplesofhepatic automatically ofhepaticci玎hosis，classi矗ed was for and cirdlosiswereinduced．Therateofidenti丘cationwas obtained，diagnostic decompensatcd principlescompeIlsated accuracy 84．6％．The resultsshowedthat for and ci玎110siscouldbe experimental diagnosticprinciplescompensateddecompensated d|．awn丘om wi吐l ci玎hosis calculation． decision一仃ee patientshepatic by words：Decision Key bree；Livermepaticcimlosis；Da协miniIl