Robust probabilistic PCA with missing data and contribution analysis for outlier detection
ComputationalStatisticsandDataAnalysis53(2009)

3706–3716
ContentslistsavailableatScienceDirect
ComputationalStatisticsandDataAnalysis
journalhomepage:

doc.guandang.net
RobustprobabilisticPCAwithmissingdataandcontributionanalysisforoutlierdetection
TaoChena, ,ElaineMartinb,GaryMontagueb
a
bSchoolofChemicalandBiomedicalEngineering,NanyangTechnologicalUniversity,Singapore637459,SingaporeSchoolofChemicalEngineeringandAdvancedMaterials,NewcastleUniversity,NewcastleuponTyne,NE17RU,UK
articleinfoabstract
Principalcomponentanalysis(PCA)isawidelyadoptedmultivariatedataanalysis
technique,withinterpretationbeingestablishedonthebasisofbothclassicallinear
projectionandaprobabilitymodel(i.e.probabilisticPCA(PPCA)).RecentlyrobustPPCA
models,byusingthemultivariatet-distribution,havebeenproposedtoconsiderthe
situationwheretheremaybeoutlierswithinthedataset.Thispaperpresentsan
overviewoftherobustPPCAtechnique,andfurtherdiscussestheissueofmissingdata.
Anexpectation-maximization(EM)algorithmispresentedforthemaximumlikelihood
estimationofthemodelparametersinthepresenceofmissingdata.Whenapplying
robustPPCAforoutlierdetection,acontributionanalysismethodisproposedtoidentify
whichvariablescontributethemosttotheoccurrenceofoutliers,providingvaluable
informationregardingthesourceofoutlyingdata.Theproposedtechniqueisdemonstrated
onnumericalexamples,andtheapplicationtooutlierdetectionanddiagnosisinan
industrialfermentationprocess.
©2009ElsevierB.V.Allrightsreserved.Articlehistory:Received9March2009Accepted20March2009Availableonline31March2009
1.Introduction
Principalcomponentanalysis(PCA)(Jolliffe,2002)isageneralmultivariatestatisticalprojectiontechniquefordimensionreduction,andithasseenawidespectrumofapplicationsinvariousareas,includingexploratorydataanalysis,patternrecognition,qualitymonitoringandcontrol.ThetraditionalapproachtotheimplementationofPCAisbasedonthelinearprojectionoftheoriginaldataontoaspacewherethevarianceismaximized.Let{xn,n=1,...,N}bethed-dimensionaldataset,thenthefirststepinPCAistocomputethesamplecovariancematrix,S,oforderd×d.TheeigenvectorswjandeigenvaluesλjofSarethencalculated(j=1,...,d).Byretainingthoseeigenvectorscorrespondingtotheqlargesteigenvalues,theq-dimensionalPCAscorevectors,tn,arecalculatedas:tn=WT(xn µ),whereW=(w1,...,wq),andµisthemeanofthedataset.Thereforetheoriginaldatacanberepresentedasalinearcombinationofthescoresplusanoisevector:xn=Wtn+µ+en.
MorerecentlyTippingandBishop(1999b)proposedaprobabilisticformulationofPCA(PPCA)fromtheperspectiveofaGaussianlatentvariablemodel.ThePPCAmodelisrealizedbyspecifyingaGaussiannoisemodele~G(0,σ2I),whichimpliesthattheconditionaldistributionofdatagivenPCAscoresis:x|t~G(Wt+µ,σ2I).ByadoptingapriorGaussiandistributionforthePCAscorevector,t~G(0,I),themarginaldistributionofthedataxisalsoshowntobeGaussian:x~G(µ,WWT+σ2I).PPCAdegeneratesintotraditionalPCAifσ2→0.WithinthePPCAframework,theprincipalcomponentsareessentiallythemaximumlikelihoodestimatesofthemodelparameters,whichcanbeimplementedusingCorrespondingauthor.Tel.:+6565138267;fax:+6567947553.
E-mailaddress:chentao@ntu.edu.sg(T.Chen).
0167-9473/$–seefrontmatter©2009ElsevierB.V.Allrightsreserved.
doi:10.1016/j.csda.2009.03.014
Robust probabilistic PCA... 11页 免费喜欢此文档... and multi-group modeling with missing data隐藏>...standard option in packaged data- analysis programs...
Robust probabilistic PCA... 11页 免费 Robust ... robust probabilistic canonical correlation analysis....and PCCA in order to handle missing data or to...
Robust probabilistic PCA... 11页 免费 Rates and Interest Rates... 31页... Cross-sectional Analysis of Longitudinal Data with Missing Values in the ...
Cross-sectional Analysis... 11页 免费 Robust probabilistic PCA... 11页 ... VIDEO IMAGE SEQUENCE ANALYSIS ESTIMATING MISSING DATA AND SEGMENTING MULTIPLE ...
Robust probabilistic P... 11页 免费 Data ...(NNs), principal component analysis (PCA) and ...(3) show that each gene contribution is weighted...
Robust probabilistic PCA... 11页 免费 A Robust Algorithm for P... 暂无... Relative Robustness, Probabilistic Robustness Analysis, Relaxation Method, Worst...
My contribution ? My future work AAM is a very hot topic (cited 2249) ... PCA with missing data (modeling), and a robust Error measure (training)...
Robust probabilistic PCA... 11页 免费 Robust machine fault det... 暂无...Robust Joint Analysis with Data Fusion in Two-Stage Quantitative Trait Genome...
Robust regression for da... 暂无评价 7页 免费 Robust probabilistic PCA.....Principal Component Analysis, Canonical Correlation Analysis and Factor Analysis ...
Analysis [1]) and PM [4] (Probabilistic ...the dimensionality of the intermediate PCA subspace... will be more robust to misregistration errors, ...
热门文档