简介:Itisnotunusualforthelevelofamonthlyeconomictimeseries,suchasindustrialproduction,retailandwholesalesales,monetaryaggregates,telephonecallsorroadaccidents,tobeinfluencedbycalendareffects.Sucheffectsarisewhenchangesoccurinthelevelofactivityresultingfromdifferencesinthecompositionofcalendarbetweenyears.Thetwomainsourcesofcalendareffectsaretradingdayvariationsandmovingfestivals.Ignoringsuchcalendareffectswillleadtosubstantialdistortionsintheidentificationstageoftimeseriesmodeling.Therefore,itismandatorytointroducecalendareffects,whentheyarepresentinatimeseries,asthecomponentofthemodelwhichonewantstoestimate.
简介:Thispaperanalyzesthedynamicsofnonlinearmultivariatetimeseriesmodelsthatisrepresentedbygeneralizedimpulseresponsefunctionsandasymmetricfunctions.Weillustratethemeasuresofshockpersistencesandasymmetriceffectsofshocksderivedfromthegeneralizedimpulseresponsefunctionsandasymmetricfunctioninbivariatesmoothtransitionregressionmodels.TheempiricalworkinvestigatesabivariatesmoothtransitionmodelofUSGDPandtheunemploymentrate.
简介:AninformationtheorymethodisproposedtotesttheGrangercausalityandcontemporaneousconditionalindependenceinGrangercausalitygraphmodels.Inthegraphs,thevertexsetdenotesthecomponentseriesofthemultivariatetimeseries,andthedirectededgesdenotecausaldependence,whiletheundirectededgesreflecttheinstantaneousdependence.Thepresenceoftheedgesismeasuredbyastatisticsbasedonconditionalmutualinformationandtestedbyapermutationprocedure.Furthermore,fortheexistedrelations,astatisticsbasedonthedifferencebetweengeneralconditionalmutualinformationandlinearconditionalmutualinformationisproposedtotestthenonlinearity.Thesignificanceofthenonlinearteststatisticsisdeterminedbyabootstrapmethodbasedonsurrogatedata.Weinvestigatethefinitesamplebehavioroftheprocedurethroughsimulationtimeserieswithdifferentdependencestructures,includinglinearandnonlinearrelations.
简介:Detectionandclarificationofcause-effectrelationshipsamongvariablesisanimportantproblemintimeseriesanalysis.Thispaperprovidesamethodthatemploysbothmutualinformationandconditionalmutualinformationtoidentifythecausalstructureofmultivariatetimeseriescausalgraphicalmodels.Athree-stepprocedureisdevelopedtolearnthecontemporaneousandthelaggedcausalrelationshipsoftimeseriescausalgraphs.Contrarytoconventionalconstraint-basedalgorithm,theproposedalgorithmdoesnotinvolveanyspecialkindsofdistributionandisnonparametric.Thesepropertiesareespeciallyappealingforinferenceoftimeseriescausalgraphswhenthepriorknowledgeaboutthedatamodelisnotavailable.Simulationsandcaseanalysisdemonstratetheeffectivenessofthemethod.
简介:Anewapproachtostudytheevolutioncomplexityofcellularautomataisproposedandexplainedthoroughlybyanexampleofelementarycellularautomatonofrule56.Usingthetoolsofdistinctexcludedblocks,computationalsearchandsymbolicdynamics,themathematicalstructureunderlyingthetimeseriesgeneratedfromtheelementarycellularautomatonofrule56isanalyzedanditscomplexityisdetermined,inwhichtheDycklanguageandCatalannumbersemergenaturally.
简介:
简介:Theareaofwellirrigationricebecamemoreandmore,sothecrisisofgroundwaterappeared.Makingthemostofrainfallisanavailabilitymethodinwatersavingirrigation,increasingwatertemperatureandraisingyield.Justbasedonthis,throughapplyingthetheoryofseasonrandomtimeseries,accordingtothedataofaveragemonthlyrainfall(1981-1999),theauthorsbuilduptheforecastingmodelintheareaofwellirrigationriceintheSanjiangPlain.Throughcontrastingwithpracticalvalue,themodelhasgoodeffect.So,itcanbeusedinwaterirrigationmanagement.
简介:ConsiderasemiparametricregressionmodelwithlineartimeserieserrorsYk=x'kβ+g(tk)+εk,1≤k≤n,whereYk'sareresponses,xk=(xk1,xk2,…,xkp)'andtk∈T包含Rarefixeddesignpoints,β=(β1,β2,…,βp)'isanunknownparametervector,g(·)isanunknonwnboundedreal-valuedfunctiondefinedonacompactsubsetTofthereallineR,andεkisalinearprocessgivenbyεk=∑j=0^∞ψjek-j,ψ0=1,where∑j=0^∞|ψj|<∞,andej,j=0,±1,±2,…,arei.i.d.randomvariables.Inthispaperweestablishtheasymptoticnormalityoftheleastsquaresestimatorofβ,asmoothestimatorofg(·),andestimatorsoftheautocovarianceandautocorrelationfunctionsofthelinearprocessεk.