1. Repository of OAI
  2. ΑΠΟΘΕΤΗΡΙΟ "ΟΛΥΜΠΙΑΣ"
  3. Σχολή Οικονομικών και Διοικητικών Επιστημών
  4. Τμήμα Οικονομικών Επιστημών
  5. Διδακτορικές Διατριβές - ΟΕ
Ελληνικά   English  
Please use this identifier to cite or link to this item: https://olympias.lib.uoi.gr/jspui/handle/123456789/39225
Full metadata record
DC FieldValueLanguage
dc.contributor.authorTatsi, Dimitraen
dc.date.accessioned2025-07-21T07:31:03Z-
dc.date.available2025-07-21T07:31:03Z-
dc.identifier.urihttps://olympias.lib.uoi.gr/jspui/handle/123456789/39225-
dc.rightsDefault License-
dc.subjectARMA(1,1)en
dc.subjectCornish-Fisher correctionsen
dc.subjectEdgeworth approximationen
dc.subjectMonte Carlo simulationen
dc.subjectRefined asymptoticsen
dc.subjectSmall sample size correctionsen
dc.subjectt and F econometric testsen
dc.subjectTime seriesen
dc.subjectMachine learningen
dc.titleSmall sample size corrections of the t- and F-econometric tests in the Linear Models with ARMA(1,1) disturbancesen
dc.typedoctoralThesisen
heal.typedoctoralThesisel
heal.type.enDoctoral thesisen
heal.type.elΔιδακτορική διατριβήel
heal.dateAvailable2025-07-21T07:32:03Z-
heal.languageenel
heal.accessfreeel
heal.recordProviderΠανεπιστήμιο Ιωαννίνων. Σχολή Οικονομικών και Διοικητικών Επιστημώνel
heal.publicationDate2025-06-26-
heal.abstractUsing refined asymptotic techniques, we derived small-sample size corrections for the t and F econometric tests in the linear regression model with ARMA(1,1) errors. These size corrections are based on the Edgeworth-corrected critical values and on the Cornish-Fisher-corrected test statistics. In particular, the size correction of the t-test can be derived from the normal or Student-t approximations. Moreover, the small-sample size corrections for the Wald and F tests can be derived from the χ2 and F approximations, respectively. Given that the Edgeworth and Cornish-Fisher corrections have an error of order O(T −3/2), where T is the sample size, the relative performance of these corrections can be investigated only by means of simulation experiments. In the context of the linear regression model with ARMA(1,1) errors, the simulation experiment conducted in this thesis seems to confirm the theoretical advantages of the Cornish-Fisher corrections. In almost all cases, the Edgeworth and Cornish-Fisher size corrections seem to improve the small-sample null rejection probabilities of the corrected t and F tests relative to the corresponding uncorrected tests.en
heal.advisorNameSymeonides, Spyridonen
heal.committeeMemberNameSymeonides, Spyridonen
heal.committeeMemberNameΣυμεωνίδης, Σπυρίδωνel
heal.committeeMemberNameVenetis, Ioannisen
heal.committeeMemberNameΒενέτης, Ιωάννηςel
heal.committeeMemberNameSimos, Theodoreen
heal.committeeMemberNameΣίμος, Θεόδωροςel
heal.committeeMemberNameSalamaliki, Paraskevien
heal.committeeMemberNameΣαλαμαλίκη, Παρασκευήel
heal.committeeMemberNameKarpetis, Christosen
heal.committeeMemberNameΚαρπέτης, Χρήστοςel
heal.committeeMemberNameKaraganis, Anastassiosen
heal.committeeMemberNameΚαραγιάννης, Αναστάσιοςel
heal.committeeMemberNameGoletsis, Yorgosen
heal.committeeMemberNameΓκωλέτσης, Γιώργοςel
heal.academicPublisherΠανεπιστήμιο Ιωαννίνων. Σχολή Οικονομικών και Διοικητικών Επιστημών. Τμήμα Οικονομικών Επιστημώνel
heal.academicPublisherIDuoiel
heal.numberOfPages257el
heal.fullTextAvailabilitytrue-
Appears in Collections:Διδακτορικές Διατριβές - ΟΕ

Show simple item record
Files in This Item:
File Description SizeFormat 
Dimitra Tatsi.pdf3.91 MBAdobe PDFView/Open
Show simple item record  


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.