ebook img

e gn ah C sp ih sn oi tal er ev itc id er p PDF

37 Pages·2017·5.09 MB·Norwegian
by  
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 e gn ah C sp ih sn oi tal er ev itc id er p

DDeDepepaparartrmtmtmeenentn totofo fIfnInIfnofofrormrmmaataitoitoinon naanandnd dSSeSerervrvivciciece eMMMaanananagagegememmeenentn tt TThThihsisidsdidsississeserertartattaitoitoinonneenendndadavavovororsrstsot oteoexexpxplpololrorereteht htehefiefiefieldeldld AAA OOO CCChhhaaannngggeee---pppoooiiinnnttt aaannnaaalllyyysssiiisss iiinnn oofofcfchchahanangngege-ep--popoioninitntatanananalaylylsysissisiisni ninpprprerededidcicticitvitvievee -otlaDD-otlaDD-otlaDD GaglG aglGagl rthrtehrteliheasliasltipasitpoitrpoirnoornobsnobshlbsehlihelmpiempismp s.a s.aIn .aItndItnpdtpdarpare nare sne usensuenmuenmtnmbstbstedbserd ierdfiorffio efffoeffrpefreporepnoesnotssntssaitbsais bailspbelspeelpeecectcststosofoff 412/02412/02412/02 kiksrokiksrokiksro ppprrreeedddiiiccctttiiivvveee rrreeelllaaatttiiiooonnnssshhhiiipppsss ssosoloulultuitoitoinonsnsosofoftfht htohososese.e.. 111 hhh 888 CCC hhh aaa nnn ggg eee --- ppp ooo OOOlglglaga aGGGoororsrskskikkikihkh h ninini ttt aa a nnn aaa ylylyl sss sisisi nini ni pp p rrr eee ddd cicici ttt vivivi eee rr r eee alalal ttt oioioi nnn sss hhh pipipi sss ISISBISBNBNN99797878-98-95-95252-62-60-60-80-82-82727171-41-4-(4p( pr(iprnirntientedted)d)) BBUBUSUSISNINIENESESSSS S+ + + 999 ISISBISBNBNN99797878-98-95-95252-62-60-60-80-82-82727272-12-1-(1p( pd(pdf)df)f) EECECOCONONONOMOMYMYY HHH ISISSISSNSNN17117979999-49-49-49393434(4p( pr(iprnirntientedted)d)) SSS ISISSISSNSNN17117979999-49-49-49494242(2p( pd(pdf)df)f) AARARTRT +T + + TTT FFF DDEDESESISGIGINGN N+ + + MMM AAaAaltalotlotUoUnUnivnieviveresrsirtsiytiyty AARARCRCHCHIHTITEITECECTCTUTURUREREE *G*G*G SScSchchohoooolololofofBfBuBususinsineinesesssss aAaAaA ciaciacia DwDwewDwewpwewpaw.pawar.atara.mtaralmttaelomtelno.tfeont.ifnt.ioftiof ofInfInfIonforfomrmramatiatoitonionananandndSdSeSerevrivrcviceiceMeMaManananagagegemememenentntt STSTECSTECCIECECIHECINHENNHCNNOCNEOCLEO LE+O L+O G+O G YGYY UotlU otlUotl hhh nnn ebebeb CCRCROROSOSSSSOSOVOVEVERERR vivivi +++ rerere DDODOCOCTCTOTORORARALAL L DDODOCOCTCTOTORORARALALL sss DDIDSISISSSESERERTRTATATATIOTIOINONSNS S DDIDSISISSSESERERTRTATATATIOTIOINONSNS S ytiytiyti 220201018188 Aalto University publication series DOCTORAL DISSERTATIONS 214/2018 Change-point analysis in predictive relationships Olga Gorskikh Aalto University School of Business Department of Information and Service Management Supervising professor Assistant Professor Doctor Pekka Malo, Aalto University, Finland Thesis advisor Assistant Professor Doctor Pauliina Ilmonen, Aalto University, Finland Preliminary examiners Assistant Professor Doctor Klaus Nordhausen, Vienna University of Technology, Austria Assistant Professor Doctor Germain Van Bever, Université de Namur, Balgium Opponent Assistant Professor Doctor Germain Van Bever, Université de Namur, Balgium Aalto University publication series DOCTORAL DISSERTATIONS 214/2018 © 2018 Olga Gorskikh ISBN 978-952-60-8271-4 (printed) ISBN 978-952-60-8272-1 (pdf) ISSN 1799-4934 (printed) ISSN 1799-4942 (pdf) http://urn.fi/URN:ISBN:978-952-60-8272-1 Unigrafia Oy Helsinki 2018 Finland Abstract Aalto University, P.O. Box 11000, FI-00076 Aalto www.aalto.fi Author Olga Gorskikh Name of the doctoral dissertation Change-point analysis in predictive relationships Publisher School of Business Unit Department of Information and Service Management Series Aalto University publication series DOCTORAL DISSERTATIONS 214/2018 Field of research Quantitative methods Date of the defence 23 November 2018 Language English Monograph Article dissertation Essay dissertation Abstract Data is the foundation of the Information Age. Knowing how to perform proper data analysis is essential and unavoidable for most companies today because it gives meaning to meaningless numbers and shows the hidden insights of information behaviour. From this perspective, change- point analysis is one of the most interesting and crucial fields, as it studies change detection in data structures and the way these changes affect underlying relationships. Change-point analysis has a considerably long history. However, the biggest part of the proposed techniques in this field are designed with a number of model restrictions, which significantly reduces the number of possible applications. In this Dissertation, we aim to study and develop robust approaches to solve the change detection problem in high-dimensional predictive structures. In Essay I, we develop a technique that allows to estimate the unknown number of changes in large datasets under normality assumptions. The proposed approach, called PSA (Parametric Splitting Algorithm), appears to be considerably accurate and efficient. In Essay II, we study away to extend the PSA method to nonparametric settings and test it with different artificial datasets. We describe this extension as the new algorithm NSA (Nonparametric Splitting Algorithm), which solves the change detection problem in a robust manner. In Essay III, we continue considering the same problem and present the new method NDP (Nonparametric Dynamic Programming) along with the proofs of its consistency. We test NDP against NSA and other baselines and conclude that, although NDP has a higher accuracy, NSA is still more preferred due to its computational efficiency. Finally, we apply NSA to news analytics to study the financial crisis of 2006–2009. Taken all together, the Essays in this Dissertation present the continuous development of the ideas towards finding a robust solution for structural changepoint detection problems in predictive relationships. Keywords data analysis, structural change, time-series, predictive relationships ISBN (printed) 978-952-60-8271-4 ISBN (pdf) 978-952-60-8272-1 ISSN (printed) 1799-4934 ISSN (pdf) 1799-4942 Location of publisher Helsinki Location of printing Helsinki Year 2018 Pages 132 urn http://urn.fi/URN:ISBN:978-952-60-8272-1 Preface "IwillnevercomebacktoFinland". "Iwillalwayshatestatistics". "Economicsisnotarealscience". Theseweremyactualwordsintheyear2013. Fiveyearslater,I’mdefending aPhDinEconomicsdealingwithstatisticsatAaltoUniversity,Finland. Little didIknowthatmyviewswouldchangethatdrastically. Indeed,change-point analysisiscrucialnotonlywithrespecttonumericaldata,butwithrespectto ourlivesaswell. Alltheselife-shiftingchangeswouldhaveneverhappenedwithoutthepeo- ple by whom I’ve been surrounded for the last four years. People like Pekka Malo. Pekka,beingthemostsupportivesupervisoronecouldeverwishfor,has taughtmetonevergiveuponmydreamsandgoals. TogetherwithPauliina Ilmonen,anotheroutstandingsupervisorIwashappytohaveguidingme,they notonlyhelpedmetobecomeabetterscientistbutalsomademefeelwelcomed inFinnishsociety. ThisDissertationwouldhaveneverseenthelightofdaywithouthelpfromthe personwhohasbeenaconstantinspirationthroughmywholelife–mymother. Mom,youhavebeenthereformeduringmyfirststepsasachildandmyfirst stepsasascientist. Thankyou. Iwouldalsoliketothankpeoplewhotaughtmetoseethebeautyofscience whileIwasstudyingattheNationalResearchNuclearUniversityMEPhI.These peopleareValeriiGalkinandSergeyErmakov(headsoftheAppliedMathemat- ics Department at the time of my studies). Additionally, with the help of my formersupervisor,AndreiAndrianov,Irealizedthatscientificresearchcanbea fruitfulprocessthatletsyouenjoyyourwork. Abunchwhodeservesaspecialplaceinthisprefacearemyfriends. Notonly becauseIlovethem,butbecausetheywon’ttalktomeifIamnottomention 1 Preface themhere. AntonFrantsev,thanksformakingmydaysinAcademiafulloffun and joy. Victoria Budarina, as you have always believed in me, I will always believe in you and our friendship, which began when we were only six years old. Andofcoursemydearestfriendsandsisters,NinaGorskikhandGalina Gorskikh: yourhard-workingnaturehasalwaysbeenaninspirationforme. Iamatrulyluckypersonasatthebeginningofmypathtowardsthisachieve- mentIhavemetthebestpersonofmylife–mybelovedhusbandPavelSelezniov. Thankyouforyoursupport,yourlove,andyourpatience. Iwillalwayscherish it. ThisDissertationisyoursaswell. Helsinki,October31,2018, OlgaGorskikh 2 Contents Preface 1 Contents 3 ListofPublications 5 Author’sContribution 7 1. Introduction 9 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2 Objectivesandresearchproblem . . . . . . . . . . . . . . . . . . 11 2. Change-pointanalysis 13 2.1 Problemsetup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2 Useofenergydistanceinstructuralchangedetection . . . . . . 14 2.3 Overviewoftheliterature . . . . . . . . . . . . . . . . . . . . . . 16 2.3.1 Parametrictechniques. . . . . . . . . . . . . . . . . . . 17 2.3.2 Nonparametrictechniques . . . . . . . . . . . . . . . . 19 2.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.4.1 Inflationmodelling . . . . . . . . . . . . . . . . . . . . . 20 2.4.2 Stockreturnmodelling . . . . . . . . . . . . . . . . . . 21 2.4.3 Wagegrowthmodelling . . . . . . . . . . . . . . . . . . 21 2.4.4 Bioinformaticalmodelling . . . . . . . . . . . . . . . . 22 2.4.5 Climatologicalmodelling . . . . . . . . . . . . . . . . . 22 2.4.6 Financialnewsanalytics . . . . . . . . . . . . . . . . . 22 2.4.7 Possiblefutureapplications . . . . . . . . . . . . . . . 23 3. Summaryoforiginalarticles 25 3.1 EssayI:SplittingAlgorithmforDetectingStructuralChanges inPredictiveRelationships . . . . . . . . . . . . . . . . . . . . . . 25 3.2 Essay II: Nonparametric Splitting Algorithm for Detecting StructuralChangesinPredictiveRelationships . . . . . . . . . 26 3 Contents 3.3 Essay III: Non-parametric Structural Change Detection in MultivariateSystems . . . . . . . . . . . . . . . . . . . . . . . . . 27 References 29 Publications 33 4 List of Publications Thisthesisconsistsofanoverviewandofthefollowingpublicationswhichare referredtointhetextbytheirRomannumerals. I OlgaGorskikh. SplittingAlgorithmforDetectingStructuralChangesinPre- dictiveRelationships. InICDM2016,NewYork,405-419,andotherdetailed information,June2016. II Olga Gorskikh, Pekka Malo, Pauliina Ilmonen. Nonparametric Splitting AlgorithmforDetectingStructuralChangesinPredictiveRelationships. In ICCDA2017,Lakeland,FL,USA,143-149,May2017. III Pekka Malo, Lauri Vitasaari, Olga Gorskikh, Pauliina Ilmonen. Non- parametricStructuralChangeDetectioninMultivariateSystems. Submitted toJournaloftheAmericanStatisticalAssociation,May2018. 5

Description:
h. C id er p h kiksr o. G a gl. O s pi hs n oit al er e vitci d er p ni sis yl ana t ni o p- e g na h. C yti sr evi n. U otl a. A. 8. 1. 0. 2 eg na h. C tci de rp. O h kiksr.
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.