Essays on the problem of distinguishing between long memory and nonlinear time series [Elektronische Ressource] / von Heri Kuswanto

Heri Kuswanto

Informations

• Publié par: gottfried_wilhelm_leibniz_universitat_hannover
• Publié le: 01 janvier 2009
• Nombre de lectures: 10
• Langue: English

Extrait

KuswEssaderyszuronpthe03.problemademiscof-distinguishingDissertationbgebet(IndonesienwdeseenGradeslonghaftswissencmemoryctorand-nonlinearontimetoseriesamVinon2009derErlangungWirtscakhhenaftDoktorswiWirtscssenschaftenhaftlicDohenrerumFoliticarumakult?tgenehmigtedervGottfriedHeriWilhelmanLe(M.Sc.)iorenbniz26.Univ1982ersit?tGresikHanno)verHendrikReferenttag:Prof.Prof.enesDr.Promotion:Philipp:SibbDr.eHakrtsenTKderorr25.06.2009eferenoABSTRAreCTpLongismofeimyporyeandisnonlievnecarwtimeproblemseriesvidedhainvhangeelongbvothtestbindicateseendetectextensiv,elytestappliedandinoem-underpiricalvstudieseonthetheMoreobusinesstestcycleurthermore,andnewotherhaptermacrobasedeconomicpritimeparametersseriesniteleadingulationtohasdierenandtineconomicords:implications.agThelopionsapproacdissertationdconsidersaldtheercomeproblemrestrictedofeconfusionNewbareetinwulationeentolongpmemorysampleander,seyvrealeraldata.nonlineareproosecesses.againstChapterin1Thedemonstrateselopinabil-theitancyiploflongstandardaggregation.methoerformancedssizetothroughdistinguish.btheetowweenabletheselongtermanwreturns.olongphedirected-Wnomgaedevna.theThebasedanalysistisodonehes,thuseroiucted-Wgtohvinthetensivofeparametersimalternativulationhstudyothesis..criticalChapteralues2prodiscussesandttensivheoreticalsimexplanationsstudyabdoneoutassessthetestsourceserformanceofnitethesize.cvonfwusion.applThethefoltoloexcwingratetFwwoprcphaptersaintesttrospuriousducememorynewctests4.ttestodevdistinguishedbonetinwarieenelongncmemoryeandestimatedsevmemoryeraltoproThecesses.pChapterin3samplecisonsialuateddesimrsstudyaItnewthattesttesttogodistinguishdboeter,wiseentolongspuriousmemorymemoryandGESTstoARknonliKeywnespuriousarimemorytnonlinear,iald,egres.tWon.ebKURZFGed?cASSUNGKapitelEineeilangfristigehprobAbh?ngiggesckann.eitsstrukturesc(sog.e"langeonGed?cvhelt.tnis"),hundwicnicahbthelineareest-PZeitreihenwirwurdenhinausausf?hrlunecicdeshtnisses"inzwiscempiriscundhenhiedenStudienun?bestererwKumonjunkturzyklenrsoAlternativhwieer-anderenundmahgef?hrt,krboh?tzen.?kf?ronomischselkursenhenkZeitreihenTangewGed?cendetestundvf?hrendeszununmittersceinemhiedlict-henAR?kunonomiscerdenhenvImplikenationen.elteDieaufvundorliegendeeinenDissertationTbProblemefasstPsicmethdemitothesederEsProb-neuelematikertederVinerwulationecdiehslungf?rzwiscnztenhenumfangeinemw"langenTGed?catenhWtnis"Dar?bundtvwirersceinenhiedenenumnicteshtnis"tlinearenDerProzessen.aufInderKapitelv1PwirdGed?cgezeigt,aggregationdasswiceseingef?hrt,mitdemStandardmethohenden"langennichhnis"tESTm?glicNichtlinearit?tenist,tersczwiscwhenkdiesenDerboneidensPh?nomenentzuk-unTterscbasiertheiden.zwDieAns?tzenAnalysevwirdendetmitdirected-WHilfeldeinerest,indastensivbenhr?nkterSimaulationadurcerhgef?hrt.eiImrzwypeitenzuKapiteleseitigen.wwerdendentheoretisckritischeWErkl?rungengeliefertf?reinedietensivUrsacSimhedurcderumVTerwerformanceeceinenhselungegrediskutiert.ImSticzweneiabzuscletzteAu?erdemKapitelendenwdiesenerdenestneueDTvestsrealenvecorgestellt,an.miterdenenenzwiscwichenelneinemin"langen4Ged?cneuenhesttnis"einundhv"langeserschhiedenenzuProzessenerizieren.unTterscwurdehiedeBasisnPrinzipswInerdearianznonkh?tztenann.arameternIn"langenKapitelh3zuwirdeeintneuerkTesttTRenditenest-PherformancehlagwderderbhegrenztenaufzudecStichhprobaggregation.engr??eist,wird"langeanhandineinerherSimen.ulationlangeaus-nicgewdirected-Wertet.inDabLageeiunecwirdtedeutlicGed?ch,tnisse"dassdenderdeutscTAktienestkeinSchohes?rter:PGed?cotentnis,tialhhalinear,tald,und.T.ABLE.OF.CONTENTSestingP.age.Atcc.kno.wledgmenprot30...deling.........54.............of...GPH.nonlinear.........8.....EST.....mo............iiiductionChapter.1:.Summary....cesses.....bi...Spurious.....ark.......mo...............53.e...duction..........1.Chapter.2:.Can4.2.1standard.tests.discriminate.bARet.w.een.commonmononlinearIntime.series.mo.de.l.s.and.long.memory?3.2.memory.........3.3.memory.of.....33.memory.cesses.......3.4.14v2.1moIn.tro.duction......3.4.2...............Conclusion.................4:.long.ory.nonlinearities...In..............4.2.24.2Long.memory.,.GPH.estimator.and.rescaled.v.ariance.testong.............4.2.26del2.3.Nonlinear.time.se.ri.e.s58models"dels3.1.tro.................................30.Characteristic.long.pro....8.2.4.T.esting.for.long.memory31.Mo.long.and.as.the.estimator.........3.4.long.in.pro...............36.M.o.swi11hing2.5delsT.esting.linearit.y..............37.Threshold.dels.........................4.3.5..............21.2.6.Conclusion................Chapter.T.for.m.m.against.AR.......4.1.tro............................26.App.endix54.The.del.................................56.L.memory.cess.......................57.EST.mo27.Chapter.3:.A.study.on."spurious.long.memory.i.n.nonli.n.ear.timeiseries.4.3.T.esting.long.memory.against.EST.AR....Main.............................5.4..59Conclusion4.3.1.T.est.statistic.and.limit.distribution......tro.................Sim......60.4.3.2.Fi.rst.approac.h........endix.............................5.1..............62.4.3.392Second.approac.h............94...................application.............106....63.4.4.Mon.te.Carlo....110...................References...................ii...................9266In4.5ductionEmpirical.application..............................5.2.result......................75.4.6.Conclusions......5.3.ulation.................................98.Empirical....................79.App.endix....5.5...................................App................................80.Chapter.5:112A.new.simple.test.against.spurious.long.memory.using.temp.oral.aggregation................115..studyAforCKNOholarship.WLEDGMENTaFirstlyA,wifeIcouldweoulddgelikyeistosonthankencouragemenmiiiyandsupofervisoracProf.ortDr.ServicePhilippofSibbyertsenbforbhismexcellenuingtcsuppeort,viencouragemenu-topandvironmengenerosithyIofwltimenancialothevExcerAAD)thealastPhDthree,ystears.mIvomwvean"tofamilyhimconsomeortvwithouterythisvhaaluableeensuggestionsprowhicdinghstimarelatingnocowerativpartenoftthisresearcthesis.andI.wgratefullyouldknoalsoeliktheesuppthankofProf.GermanDr.cademicHendrikhangeHak(DenebswandyProf.aDr.scLukFinallyasmMenkhogreatefordebttheirtokindnessyandelosuppedort"Nora",duringymeloyedPh"RyDandstudyyhere.,Thankstheiralsotintosuppallandthet,memwhibhersthesisofnotthevInstitutebofwritten.Statistics,forth1asChaptertheoretical1MonSUMMARWYproThisrdissertationedealsChapterwithofthenproblemmemoryoforconfusingmobnonlinearetewweenconfusionlongviormemorytheandtesomeiscommoneclassesmethooffornonlinearteststimeaseriesshoproundercess.cessItoreerties.ssenastiallyandconthetainssomeofthefourproconcotributionsandtoprotheTheproblem.theseThethedissertationconditions.pursuitsfromalsoythreesucmaGPHjormemorygoals.earitOneiisortoecicassessWtbheCarloabilitconditions,ygeneratingofhaexistingeithermetholongdsorytoproblemdistinguishknolongbmemorylongfromprotheidennoofnylex-ineardescribprobcesses.fTheinsecondau-goalarianceiscorrelationtoortidenbtifysomethesimsourcesofofmemorytheproconfusion,bandoftheinlastnisdistinguishedtolongdevbelopstandardtestsdswhichhtheareestimatorablethetoparameterdistinguishlin-longymemoryeandthenonlineargeneralproagainstcespsses.nonlinearThedel.outlineeowfyeactehthatccertainhaptertheisdatadescribproedcanasvfollomisleadingws.stationaryThenon-stationarysecondmcmhapterpropdealsThiswithistheusrstwngoal.confusionThroughetaneeninmemorytensivnonlinearecesses.sim3ulationtiesstudysources,thewbeexploringshotheoreticalwplanations.thatespeecicasymptoticnonlinearehatimeosethericessetermssthemotodelsvsucandhautoasfunction,SETsuppAR,theLSTevidencesAR,yESTvidingARMonandCarloMarkulations.oexistencevloswitcghing,inwhicnonlinearhcessesareinducedcommonyinnatureeconometricthepracticecesscancertainhardlIybγ
propaddition,parameter.theyGPHtheestimatorcounitselfstatisticisulationsbiased.realMostfactofpropthetheconaggregation,tributionsaintothethesis,dissertationatranspireainPPPtthisheetaniosforthresultandtestfthincforhapter.tInofcwhapterwith4,Japanwseveofdeelopingvableelopproblem.atWholdsaldaretapplyypBaieantestprotowithdistinguishdiscussesbmemoryetedwlongeenthelonginmemorylandtheESTalueARinonlinearittheysamples,bfromyousingesaIndeicasesrehangected-Welopalddstatistictries.totestsoprovtoercomeenthenonlinearproblemwoflikrestricteddevparameterstheundertotheWalternativae.bTheandtesttoisneglectedderivlongedandfromconsistentFinallywconewbasicspuriousmoThedeldevsponecications,ariancewhereparametersthebinedrstofiscthelongstandardusingmocaldelebasedtakonumancomauxiliaryaggregatedregressions.andwthepsecondinalloiswslongtheproparametert2er.vtoapplicationappagainstearYEN.asthisawnconsideruieralsancincludingeexcparameterratesindevtheedtransitionnfunction.devAcounsimTheulationosedstudyareindicatestothatvidebsolutionoththeapproacGivhestheleadthattoadjustmenteststowithardsgomoreoelydinsizeelopingandtries,ptestsoablewcaptureerphenomena.propeealsortiestesttooseddistinguishybllieetKapw(2008)eendetectstationaryylnonlinearitonginmemorymemoryandcessesESTtheAR.isMoreotvours.er,,thelastsecondhapterapproacahsimpleisagainstsholongwn.totesthaisvelopebasedmorethepvoofwmemoryer.toAscomanwithempiricalideaapplication,testingwaehangeapplythethememorytestBytothetheoproblemWhiaddressetldestimator,bstatisticyesCheungmaximandvLaiamong(2001).binationsTheypairedfacedserdicultieseinSimdistishonguishingthatlongtestmemoryerformsandellnonlinearnitemeanandrevableersiondistinguishinmemorythespuriouscasecessesofexcellenbilateralprealwexcMoreohangeer,rempiricalagivte