165 Pages
English

Model selection procedure with familywise error rate control for binomial order-restricted problems [Elektronische Ressource] / von Xuefei Mi

-

Gain access to the library to view online
Learn more

Description

orMobaudelakSgenehmigteelectionHannoProeinescedureDr.withM.Sc.FinaErlangungmheniderlywisehaftenErrorhort.RatevConMitrolamforPBinomialer,Order-RestricteddesProblemsademiscVGradesonDoktorsderGartenNaturwissenscwissenschaftlic-henrer.F-akult?tDissertationderonGottfriedXuefeiWilhelmgebLeibnizenUniv13.05.1980ersit?tShandong,Hanno.R.Chinavver2009zurr2WReferenagt:tProf.herDr.Promotion:LudwigalHothorneKLehmacorreferenTt:derProf.23.03.2009Dr.deiungsbZusammenfassungVMokdell-basiertewieunddieTultipleest-basiertedieMethosicdenFWERwwirerdenahlmeiasothesen,tehnsAufzurSeiteAnalyseest-gegr?ndetedeerteilungrNacDosis-aucWirkungInformationskriteriumBeziehMustungbvonerwhlagwendet.ordnDerMethoGebrauctrierthwievInformationonaordneungseingesconzenhr?nktendieHy-depeineothesenundistoneinedeeinheitlichhewMethomodtieren,ef?r,r-dieterPtenohr?nkungwerenern.v:ergr??ert,ondenhr?nktenalternativformationskriteriumendeRaumonzeneinengend.sicHiermitauf,sindman?realendherausndet.ederrunngspunkte,reinfacnhekOrdntriertunghundTeinfacMethoherauf,BaummandreiVallgemeaufbautinedasTkyptrolliert.enhvmonergleicOrdnstellen,ungsbhescerdenhr?nkungen.einWirdizierteswpr?

Subjects

Informations

Published by
Published 01 January 2009
Reads 18
Language English
Document size 1 MB

orMobaudelakSgenehmigteelectionHannoProeinescedureDr.withM.Sc.FinaErlangungmheniderlywisehaftenErrorhort.RatevConMitrolamforPBinomialer,Order-RestricteddesProblemsademiscVGradesonDoktorsderGartenNaturwissenscwissenschaftlic-henrer.F-akult?tDissertationderonGottfriedXuefeiWilhelmgebLeibnizenUniv13.05.1980ersit?tShandong,Hanno.R.Chinavver2009zurr2WReferenagt:tProf.herDr.Promotion:LudwigalHothorneKLehmacorreferenTt:derProf.23.03.2009Dr.deiungsbZusammenfassungVMokdell-basiertewieunddieTultipleest-basiertedieMethosicdenFWERwwirerdenahlmeiasothesen,tehnsAufzurSeiteAnalyseest-gegr?ndetedeerteilungrNacDosis-aucWirkungInformationskriteriumBeziehMustungbvonerwhlagwendet.ordnDerMethoGebrauctrierthwievInformationonaordneungseingesconzenhr?nktendieHy-depeineothesenundistoneinedeeinheitlichhewMethomodtieren,ef?r,r-dieterPtenohr?nkungwerenern.v:ergr??ert,ondenhr?nktenalternativformationskriteriumendeRaumonzeneinengend.sicHiermitauf,sindman?realendherausndet.ederrunngspunkte,reinfacnhekOrdntriertunghundTeinfacMethoherauf,BaummandreiVallgemeaufbautinedasTkyptrolliert.enhvmonergleicOrdnstellen,ungsbhescerdenhr?nkungen.einWirdizierteswpr?senerdendasdieseFWERzwdieeieMethoauswdenunimder?nderungspunkte-EnestimmtdecOrdnkungsproblemescundkan-trolliderenkOrndnScungsborteescmhr?nkungsproblemKvtrasttests,ergleicungseingeschen.HypDieIn-Modell-gegr?ndeteFinallyiieAbstractstrictions,Moedel-basedw(RtroloWystonnetalueal.,suitable1999)Tandselectiontest-based(MHI(DosemeciypandtheBealternativnofitoclikhou,calculating1998)omethoCriteriondsmoareordermostHothorncommonlymousedthetotheanalyzecandidatedose-respetonseandrelationship.willThewillusecalculateoftheorder-restricted"ohdifyypestimatorsothesesdisunderaSimple-ordercommonord:apprypoacancFWERhewhicdierenhFirst,increasesMipCri-otowselectierwillbullyandnarrotaryw-delsingdels.dodierenceswneenthemoalternativelemenemospace.eHerebwythe,dierencesChange-pcriticaloincont,InSimple-ordereandertting"Simple-treealsoaremaximthreeocommonintelihoyp(SLE)esinformationoforderorderhrestrictions.andInrestriction.thisConthesis,Order-restrictedwtest,ehwillconcomthepforaredmoldel-basedundermtethorestrictions.dswwithbuildtest-basedandmethoInformationdsterioninC)Change-pdooindelton.detectioneandconsiderothernorderhrestrictionothesisproblems.allTheelemenmoalternativdel-basedmomethoasdmofoSecond,cusesinformationonbthewrealtheinformationulldistance.delOnthethetaryothereside,delsthebtest-basedcalculated.metho,defobuildcusesdistributiononthesehoandwthetovbuildtoatroldistributionFWER.andordertosolvconthetrolvtheproblem,Feamilywisemoer-therorumrateeliho(FWER).dAfter(MLE)thetocomparison,likwoeestimatorswillforalsothepresencriteriontcertainare-newsucmethoasd,restrictioncalledSimple-treeMultiplederLikKeywelihoMultipleotrastdest,Thestothesis(MLInformationT),whictagesConSuitableten.ts.1.In.tro.duction261distance2.Motiv.ationse5.2.1(SCT)Dose-Resp.onse.Studies......its.................l...(CA...ds.mo...Bias...(OSAIC).....Le...Ninom5.2.1.1eA.clinicalestimatordoseofnding.study.withTan.adv.erseConev.en.ts23rateT.....hran-Armitag5.2.2.Epidemiological3.4.5case-tagescon.trolInformationstudies4.1.......k.....32.....P.mators.....y.......(NIC)..6.2.2.140Suldic.nic.k2eliklenaltand44lungcalcdancer..................3.4.2.T...........Multi.Con.(MCT)6.2.2.2.The.eect.ofCoaTge.on.the.sp.on.taneousdvabdisadvortiontestrate....28.(IC).selection.k-Leibler....6.2.3.Bio-informatics:.DNA-motif31ndtheiong..........One-sided.ORIC.......36.sets.st.........38.el..7.3.Mo.del.selection.pro39cedurea13.3.1.Order.restriction......Ak.Criterion.........Impro.maxim.o.y.term...Notic.global,.n.eliho.mators...............45.)..13.3.1.1.Single.Change-p.oin.t.order.restric.tion23.Single.trast.est.....................3.4.313p3.1.2eEpidemitrastc-orderestrestriction..................3.4.4.c.e.est.T)..............16.3.1.327SimpleA-orderanrestrictionand.an.of.metho...........4.Criterion.for.del.31.Kullbac.(KL)..............17.3.1.4.Simple.-tree.restriction4.1.1.of.Log-Li.eliho.d.....................4.1.2.AIC.and..........18.3.2.Denition.of4.1.3moartitiondelandselectione.i...................4.1.4.v.probabilit....................19.3.3.F.amilywise4.1.5erroriyrateAIC(FWER).c.on.trol..................4.1.6.aik.Information.(AIC)..........21.3.4.Previous4test4.2methovdsthe.um.eliho.d.b.p.y.......4.2.1.ation.the.lo.a.d.Lik.o.Esti-................22.3.4.1.Lik.eliho.o.d.Ratio.T.estiii(LR.iv.CONTENTSorder4.2.2.EstimaMt.ors.undermatrixSingle7.4.2Change-p.oin.t.orderandrestriction..ccuracy......49.4.2.3.Estima.t.ors.under.Epidemic-order.renitione.st.ri.ction........ction.......6.2....53.4.2.4mic-orderEstima.t.ors.under.Simple-order.restric.tion..w.........u...calculating......551244.2.5.EstimaBinotrendt.orsdeunder.Simple-treeChange-pord.e.r128restriction..................e61.4.3.Sim.ulation.study.for.comparing6.2.3gMLE,.lMLE.and.SLE6.2.4................64Multiv5MvnormT.est-based.mo.del.selectionon67.5.1.Relationship.b.et.w.een7.1.4Log-lik.eliho.o.d.Ratio.T.estities(LR.T)teredand.Multiple.ConPtrast.T.e.st.(MCT)......127.r...................91.rate(MR)...........ulation............67.5.1.16.2.1Doinistribut.i.on.of.the.log-lik6.2.2eliho.o.d.under.Single.Change-p.oin.tple-orderorder.restriction............ple-tree.............6.3.................7.107.Normal.k..68.5.27.1.1MultiplepropLog-lik.eliho.o.d.T.est7.1.2(MLalgorithmT).with.con.t.rol.of7.1.3FWERerical............73es5.2.1vnCritic.al.v.alue..7.2.time...........7.2.1.tri-diagonal.....Ce.t.........124.k...............7.473summary5.3.ORIC-lMLE,.MHIC.an.d.ML.TSinglundtestrr.order.restriction....mic-order...........(CR)..74.5.3.1.Singl.e.Change-p.oin.t.order6.1.3risclassicatione.str.i.ction..............92.Sim.study..........74.5.3.2.Epide.mic-order.restriction..........93.Singl.Change-p.t.....................93.Epide75.5.3.3.Sim.ple-order.restriction..................98.Sim........................76.5.3.4.Sim100ple-treeSimorderorderrestriction........................100.Conclusion............77.5.4.Relationship.b.et.w.een.MCT.and.ML.T101.Soft.are.7.1.ariate.Distribution.pac.age.........107.D.and.erties......79.5.5.Algebraic.space........108.M.te-Carlo.......................110.N.m.algorithm..................80.5.6.Mo113delExamplselecfortionmwith.con.trol.of.FWER.for.ML.T....122.A.and.consumption..............85.6.P.o123wProbabilerwithstudycorrelationand.sim.ulation.89.6.17.2.2Expressionsn.orthan.probabilities...................7.3.ac.age.............................125.Co.for.section89.6.1.1.Expressio.n.of.the.p.o.w.er......7.4.1.e.oin.order.e.i...............127.Epide.restriction......90.6.1.2.Correc.t.mo.del.sele.ctionrate141CONTENTS.v.7.4.38.2.1Simpl.e.order.restriction......restriction...................8.1.4.....140............129.7.4.4.Simpl.e-tree.restriction..............e-tree...........Conclusions.................n131.8.Summary.133.8.1.Solution.to.theTheprevious.examples..........141...................138.Simpl.order......133.8.1.1.Single.Change-p.oin.t.order.re8.2s.tric.tion............................141133Mai8.1.2resultsEpidemi.c-order.restriction........................8.2.2.relationship..............135.8.1.3.Simpl.e-order.restriction.....CONTENTSvir=3;s=13 r=2;s=13 r=3;s=14
H H HA A A
r=3;s=12 r=3;s=14H H
A A
simmaximNICum.andalsen.tropdel-basedyov(thealues.(vertan.ZwoldetUnderet.al.,w2005)inofisthenaligned2004).DNAossimotif..8.3.The.....asymmetric.NIC.ultaneous.ossible.v.oin.dieren.largest.mo...del...DNA.Figures.pattern........test-based.our.mark...vii..11993.1alternativSimpattern:ulationerfor.the8.1data:inPalloidels.nplottstestarekgeneratedandpropalsortions;moSpheres.arealtheb95%2.2condence.regieonsthat.......,.ab.Prop.List.l...............relationship.d,.d.metho.creations.in..15.5.1.Threeofdimension.plot.for.sim6.4ulatedthebinomiale,data5x3k=2p.w.of.....100.Sim.condence.terv86for5.2pEnlargedmothreeHeredimeensiontheplotalueforstatisticssimblaculatedpbinomialts)datatheirk=2terv.for.t.dels87ultaneously6.1TheSimvulationueofobtainedpyodelw8er.and.moWdelalsoselecdThemoalso.v.relativ.larger,alue.others..8.2(Lewin,trastsMotiftheout5yandanden2.1yofofmostthaieonelyratev.among.136.Con.for.top.pattern.the.trop.comparison.the.p.b.e..94.6.2.A.ccording.to.the.n.ull,.the.nding.rate.of.NIC137isTheacceptableof.metho.mo.metho.and.new.d.99new6.3areUnderedthebalternativ.e,asymmetric.3x3.pattern:.p.o142werOFviiiFIGURESLIST