資源描述：
《THE EFFECT OF MODEL MISSPECIFICATION ONSEMI-SUPERVISED CLASSIFICATION》由會員上傳分享，免費在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、IEEETRANSACTIONSONPATTERNANALYSISANDMACHINEINTELLIGENCE,VOL.33,NO.10,OCTOBER20112093TheEffectofModelMisspecificationonSemi-SupervisedClassificationTingYangandCareyE.Priebe,SeniorMember,IEEEAbstract—Semi-supervisedclassification—trainingbothonlabeledandunlabeledobservations—can
2、yieldimprovedperformancecomparedtotheclassifierbasedononlythelabeledobservations.Unlabeledobservationsarealwaysbeneficialtoclassificationifthemodelweassumeiscorrect.However,theymaydegradetheclassifierperformancewhenthemodelismisspecified.Intheclassicalclassificationproblemsett
3、ing,manyfactorsaffectthesemi-supervisedperformance,includingtrainingdata,modelspecification,estimationmethod,andtheclassifieritself.Forconcreteness,weconsidermaximumlikelihoodestimationinfinitemixturemodelsandtheBayesplug-inclassifier,duetotheirubiquitousnessandtractability.In
4、thisspecificsetting,weexaminetheeffectofmodelmisspecificationonsemi-supervisedclassificationperformanceandshedsomelightonwhenandwhyperformancedegradationoccurs.IndexTerms—Semi-supervisedclassification,finitemixturemodel,Bayesplug-inclassifier.?1INTRODUCTION1.1ProbabilisticMode
5、lLetbethelimitofthesupervisedMLEof,andsupthelimitoftheunsupervisedMLEof.BothandETeX;YTFXY.ThefeatureobservationXisanunsupsupLIRd-valuedrandomvariable.Thenatureoftheobserva-existundermildregularityconditions.unsuptioniscalledaclasslabel,denotedbyYandtakingvaluesForsi
6、mplicity,weconsidertwo-classclassificationinafinitesetf1;2;...;Kg.Forj?1;...;K,denotetheclassproblemsthroughoutthispaper.ThejointdensityofconditionaldistributionsbyFj?FXjY?j,andassumeweareeX;YT,feX;YTex;yT,canbewrittenasinthecontinuouscase,sotheclassconditionaldensitiesfjexist
7、.Letj?PfY?jgbeclasspriors,whichcanalsobe1fexj1T11fy?1gte11Tfexj2T11fy?2g:referredtoascomponentcoefficients.WeassumethattheclasslabelYisnotobserved,andour1.2PreviousWorkgoalistoclassifythefeatureobservationXwithsmallThequestionsofvalueandriskofsemi-supervisedlearningclassif
8、icationerrorLegT?PfgeXT6?Yg.Supposewearehasbeeninvestigatedby
