Hareketli ortalamalarla bütünleştirilmiş otoregresif süreçler (arıma) ve bazı iktisadi zaman serilerine uygulanması

Abstract

ÖZET HAREKETLİ ORTALAMALARLA BÜTÜNLEŞTİRİLMİŞ OTOREGRESİF (ARIMA) MODELLER VE BAZI İKTİSADİ ZAMAN SERİLERİNE UYGULANMASI Ebru ÖZGÜR Yüksek Lisans Tezi, Ekonometri Anabilim Dalı Danışman : Doç. Dr. H. Altan ÇABUK Temmuz 1998, 133 sayfa Bu çalışmada; tek değişkenli öngörü modellerinden, son yıllarda uygulamada yaygın olarak kullanılan Box - Jenkins metodolojisinin istatistiksel özelliklerinin teorik açıdan ele alınması amaçlanmıştır. Çalışmada, bu amaçla, zaman serileri analizinde önemli bir yer tutan fark denklemleri ile fark denklemlerinin çözüm teknikleri, tek değişkenli öngörü modellerinde kullanılan testler ve Box - Jenkins yöntemiyle model seçiminde izlenen aşamalar incelenmiştir. Bu teorik bilgiler ve kriterler göz önünde bulundurularak aylık ve yıllık olmak üzere 4 iktisadi zaman serisi modellenerek, öngörüde bulunulmuştur. Anahtar Kelimeler : Zaman Serisi, Tek Değişkenli Modeller, Box- Jenkins, ÖngörüABSTRACT AUTOREGRESSIVE INTEGRATED MOVING AVERAGE MODELS AND APPLICATIONS ON SOME ECONOMIC İTME SERIES Ebru ÖZGÜR Master Thesis, Department of Econometrics Supervisor : Doç. Dr. H. Altan ÇABUK July 1998, 133 pages The purpose of this study is to investigate theoretically the statistical characteristics of Box - Jenkins methodology which is a commonly used univariate forecasting model. Therefore, in this study, differential equation which plays an important role in time series analysis and their solution techniques, tests used in univariate forecasting models, and the phases pursued in model selection by means of Box - Jenkins method have been studied. Taking into consideration yhe theoretical information and criteria available was made by modelling four different monthly and yearly economic time series. Keywords ; Time Series, Univariate Modşls, Box - Jenkins, Forecasting

ABSTRACT AUTOREGRESSIVE INTEGRATED MOVING AVERAGE MODELS AND APPLICATIONS ON SOME ECONOMIC İTME SERIES Ebru ÖZGÜR Master Thesis, Department of Econometrics Supervisor : Doç. Dr. H. Altan ÇABUK July 1998, 133 pages The purpose of this study is to investigate theoretically the statistical characteristics of Box - Jenkins methodology which is a commonly used univariate forecasting model. Therefore, in this study, differential equation which plays an important role in time series analysis and their solution techniques, tests used in univariate forecasting models, and the phases pursued in model selection by means of Box - Jenkins method have been studied. Taking into consideration yhe theoretical information and criteria available was made by modelling four different monthly and yearly economic time series. Keywords ; Time Series, Univariate Modşls, Box - Jenkins, Forecasting