Döviz kuru volatilitesini öngörmede melez bir model: Yapay sinir ağı tabanlı egarch
Yükleniyor...
Tarih
2022
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İstanbul Okan Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Döviz kurlarında yaşanan dalgalanmalar, volatilite modellerinde yeni arayışlar yapılmasına neden olmaktadır. Bu çalışmada 2018-2019 dönemi günlük verileriyle ABD doları/Türk Lirası reel alış kuru üzerinde öngörü yapılması amaçlanmıştır. Bu bağlamda GARCH Tipi modellerin kabiliyetini artırmak amacıyla geliştirilmiş yapay sinir ağı tabanlı EGARCH modeli ile öngörü yapılmıştır. GARCH tipi modellerden GARCH (1,1), GJR-GARCH (1,1) ve EGARCH (1,1) ile modellemeler yapılmış, ilgili bilgi kriterine göre etkin model olan EGARCH (1,1) modelinden elde edilen hatalar ile ile sinir ağı modeli kurulmuştur ve melez bir öngörü modeli inşa edilmiştir. Yapılan incelemeler sonucunda uygulanan melez modelin öngörü performansının GARCH tipi modellere kıyasla daha iyi olduğu görülmüştür.
Fluctuations in exchange rates cause new searches for volatility models. In this study, it is aimed to predict the real buying rate of US Dollar / Turkish Lira with daily data of 2018-2019 period. In this context, an artificial neural network based EGARCH model was developed to increase the capability of GARCH type models. GARCH type models were modeled with GARCH (1,1), GJR-GARCH (1,1) and EGARCH (1,1). A hybrid prediction model was constructed by channeling the errors obtained from the EGARCH (1,1) model, which is the effective model according to the relevant information criteria, to the artificial neural network. As a result of the examinations, the predictive performance of the hybrid model was found to be better than the GARCH type models.
Fluctuations in exchange rates cause new searches for volatility models. In this study, it is aimed to predict the real buying rate of US Dollar / Turkish Lira with daily data of 2018-2019 period. In this context, an artificial neural network based EGARCH model was developed to increase the capability of GARCH type models. GARCH type models were modeled with GARCH (1,1), GJR-GARCH (1,1) and EGARCH (1,1). A hybrid prediction model was constructed by channeling the errors obtained from the EGARCH (1,1) model, which is the effective model according to the relevant information criteria, to the artificial neural network. As a result of the examinations, the predictive performance of the hybrid model was found to be better than the GARCH type models.
Açıklama
Anahtar Kelimeler
Volatilite, Zaman Serisi Analizi, GARCH Model, Yapay Sinir Ağı, Öngörü, Volatility, Time Series Analysis, GARCH Model, Artificial Neural Network, Forecasting
Kaynak
WoS Q Değeri
Scopus Q Değeri
Cilt
0
Sayı
659
Künye
Bezgin, M., Kaya, E. (2022). Döviz kuru volatilitesini öngörmede melez bir model: Yapay sinir ağı tabanlı egarch. Finans Politik ve Ekonomik Yorumlar Dergisi, 0(659), 115 - 133.
