ebook img

Linear Time Series with MATLAB and OCTAVE PDF

355 Pages·2019·5.031 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Linear Time Series with MATLAB and OCTAVE

Statistics and Computing Víctor Gómez Linear Time Series with MATLAB and OCTAVE QUANTLETS Statistics and Computing SeriesEditor WolfgangKarlHärdle,Humboldt-UniversitätzuBerlin,Berlin,Germany Statistics and Computing (SC) includes monographs and advanced texts on statisticalcomputingandstatisticalpackages. Moreinformationaboutthisseriesathttp://www.springer.com/series/3022 Víctor Gómez Linear Time Series with MATLAB and OCTAVE 123 VíctorGómez GeneralDirectorateofBudgets MinistryofFinanceandPublic Administrations Madrid,Spain Quantletsmay bedownloaded fromhttp://extras.springer.com or viaalink onhttp://www. springer.com/978-3-030-20789-2orwww.quantlet.orgforarepositoryofquantlets. ISSN1431-8784 ISSN2197-1706 (electronic) StatisticsandComputing ISBN978-3-030-20789-2 ISBN978-3-030-20790-8 (eBook) https://doi.org/10.1007/978-3-030-20790-8 MathematicsSubjectClassification(2010):62-01,62-02,62-04,62-07,62M15,62M10,62M20,62J05 ©SpringerNatureSwitzerlandAG2019 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthors,andtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional claimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG. Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland To mywifeMaríaC. and mydaughterBerta Preface Themotivationofthisbookistoprovidetimeseriesstudentsandresearcherswitha softwarepackagecalledSSMMATLAB,writteninMATLAB,thatwillallowthem to work with general state space models. Since many time series models used in practicecanbeputintostatespaceform,specialfunctionshavebeenwrittenforthe mostusualones,suchasmultiplicativeARIMAandVARMAmodels,cointegrated VARMA models, VARMAX models in echelon form, transfer function models, univariate structural models, like those considered by Harvey (1993, Chap. 4) or Kitagawa and Gersch (1996), and ARIMA model-based (AMB) unobserved components models (Gómez and Maravall 2001). However, if the user intends to workwithmoresophisticatedstatespacemodelsthatarenotavailableinstandard commercialpackagesfortimeseriesanalysisoreconometrics,he/shecanprogram his/herownmodelinSSMMATLABandcarryoutmodelestimation,interpolation, forecasting,andsmoothing. All the programscontainedin SSMMATLAB can also runin the free software OCTAVE platform. The series can be univariate or multivariate and the state space model can be very general. It may have time-varying system matrices, exogenousinputs,regressioneffects,incompletelyspecifiedinitialconditions,such asthosethatarisewithnonstationaryVARMAmodels,andmissingvalues.Abrief descriptionofSSMMATLABappearedinGómez(2015). Thebookprovidesaquicktheoreticaldescriptionofa widevarietyoftopicsin Time Series, togetherwith manyexamplesusing SSMMATLABto illustrate each subject. Many of the examples in the book can be found as Quantlets in Gómez (2019),whereSSMMATLABcanalso bedownloaded.Wheneverthe readerfinds an examplein which the logo followedby some name appears,thismeansthat this is a link to the correspondingquantletwith the same name where the code to producethepictureortheresultsintheexampleislocated.Inthisconnection,itis tobenotedthatinthelinktotheQuantletplatform,www.quantlet.de,onecanfind allthecontentprovidedathttps://github.com/QuantLet. ThecompanionbookGómez(2016)containsthetheoreticalmaterialneededto develop most of the functions used in SSMMATLAB. The user is encouraged to vii viii Preface consultthisbookinordertounderstandtheprogrammingofmanyofthealgorithms usedinthesefunctions. Finally, I would like to thank several people who have helped in some way to develop this project. In particular, thanks are given to Juan Varela of the Spanish Ministry of Finance, Félix Aparicio and Luis Sanguiao of the Spanish National Statistical Institute, and Riccardo Corradini of the Italian National Statistical Institute. The book is organized as follows. A quick introduction to SSMMATLAB is provided in Chap.1, where some examples are given of the most widely used models in practice. In the rest of the chapters, for each topic in time series analysis,atheoreticalintroductionisgivenfirst,followedbysomeillustrationwith SSMMATLAB. More specifically, in Chap.2, stationarity, VARMA and ARIMA models are considered. VARMAX and transfer function models are described in Chap.3. Unobserved components in univariate series is the subject of Chap.4. Chapter 5 is dedicated to spectral analysis. The computation of echelon forms for VARMAX and state space models by polynomial methods is the subject of Chap.6.InChap.7,severalexamplesofmultivariatestructuralmodelsarepresented thatcannotbehandledwithstandardcommercialpackages.CointegratedVARMA modelsareconsideredinChap.8.Chapter9showshowARIMA,VARMA,VAR- MAX, and time invariant state space models can be simulated in SSMMATLAB. The state space model is the subject of Chap.10. All the examples contained in SSMMATLABarelistedbysubjectinChap.11. Madrid,Spain VíctorGómez August2019 Software Installation ToinstallSSMMATLAB,uncompressthezipfilecontainedinthedirectoryToolbox ofthewebpageGómez(2019)andcopyitscontentsintoa directory,forexample SSMMATLAB. Apart from the files with extension m (MATLAB m-files), there shouldbe six subdirectorieswith namesDATA,RESULTS, SPEC, GRAPHS, and USMSPEC.AllthedatafilesusedinthedemosareinthesubdirectoryDATA.The subdirectoryRESULTS is where allprogramresultsare written. The subdirectory GRAPHS is used to eventually write the plots produced by the programs. In the subdirectory SPEC, you can find all specification files for the different ARIMA andtransferfunctiondemos.Thatis,eachspecificationfilecontainsinstructionsto readdata,generatestructures,etc.Finally,thesubdirectoryUSMSPEChasthesame functionasthesubdirectorySPECbutforunivariatestructuralmodels. Iftheuserdesirestoworkinadirectorydifferenttotheoneinwhichtheprogram has been installed, he/she should first add to the MATLAB path the directory whereSSMMATLABhasbeeninstalled.Then,ifthenewdirectoryinwhichthe userintendstoworkiscalledWORK,forexample,theusercanstartworkingwith SSMMATLAB in that directory without having to create any subdirectory. Some oftheprogramsusedbySSMMATLAB,forexamplethosethatdealwithARIMA, transferfunction,orunivariatestructuralmodels,canoptionallyandautomatically create two subdirectories in WORK, called RESULTS and GRAPHS, where the programresultsandplotsarewritten. AllthefunctionsinSSMMATLABhavebeenprovedtoalsorununderthefree softwareOCTAVEplatform. References Gómez,V.(2015).Ssmmatlab:Asetofmatlabprogramsforthestatisticalanalysis of state spacemodels.JournalofStatisticalSoftware,66(9),1–37.http://www. jstatsoft.org/v66/i09/ ix x SoftwareInstallation Gómez, V. (2016). Multivariate time series models with linear state space struc- ture.NewYork:Springer. Gómez,V.(2019).SSMMATLAB.https://github.com/QuantLet/ssmmatlab Gómez, V., & Maravall, A. (2001). Seasonal adjustment and signal extraction in economictime series.In D.Peña,G. C. Tiao,&R. S. Tsay (Eds.),A coursein timeseriesanalysis(chap.8).NewYork:Wiley. Harvey,A. C. (1993).Time series models(2nded.).HemelHempstead:Harvester Wheatsheaf. Kitagawa,G.,&Gersch,W.(1996).Smoothnesspriorsanalysisoftimeseries.New York:Springer.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.