Finansal piyasalarda oluşan belirsizliğin ve riskin giderilmesi amacıyla geliştirilmiş olan türev piyasalarda, piyasaya duyulan güven, doğru bilginin piyasaya dahil olan tüm unsurlara aynı anda ulaşması sayesinde piyasanın etkin olarak işlemesi durumunda gerçekleşebilmektedir. Böylece, geçmiş dönem fiyat hareketlerinden yararlanarak gelecek döneme ilişkin öngörüler yapmak mümkün olmamaktadır. Bu bağlamda çalışmada öncelikle, Türkiye’de faaliyet gösteren Vadeli İşlem ve Opsiyon Piyasası’nın etkinliği; Genişletilmiş (Augmented) Dickey-Fuller (ADF), Phillips-Perron (PP) ve Kwiatkowski vd. (KPSS) doğrusal birim kök testleri ve Kapetanios vd. (KSS) doğrusal olmayan birim kök testi uygulanarak sınanmıştır. Rassal yürüyüş sergilemediğine karar verilen seriler sebebiyle piyasanın etkin olmadığı sonucuna ulaşılmıştır. Ardından, Vadeli İşlem ve Opsiyon Piyasası’nda işlem gören TL/Dolar ve Bist- 30 sözleşmelerinin gün sonu uzlaşma fiyatının öngörüsünde en yüksek performansı gösteren yöntemin belirlenmesi amaçlanmıştır. Bu amaçla, Borsa İstanbul A.Ş’den temin edilen ve 04.02.2005 – 31.12.2015 tarihleri arasını kapsayan veriler kullanılmıştır. Analiz bulgularına göre, TL/Dolar sözleşme serisi için ARMA(4,4) modeli, RBF-1-B-L yapay sinir ağı modeli ve ARCH(1) modeline kıyasla daha yüksek öngörü performansı gösterirken, Bist- 30 sözleşme serisi için ise TDNN 1-B-L yapay sinir ağı modeli, ARMA(4,5) ve ARCH(1) modeline kıyasla daha yüksek öngörü performansı gösteren model olmuştur.
Derivative markets developed to eliminate uncertainty and risk arising from financial markets can make predictions about the future by using past price movements in case the market is not effective. In this context, in this study, firstly, the effectiveness of the Turkish Derivatives Market was tested by applying the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP) and Kwiatkowski et al. (KPSS) linear unit root tests and capetanios et al. The nonlinear unit root test. As a result of all unit root tests, it was concluded that the series did not show random walk, so that the market was not effective. Then, the method that shows the highest performance is tried to be determined when forecasting the end of day settlement price of the USD/USD and Bist-30 contracts which is traded in the Derivatives Market. For this purpose, the forecasting results produced by the time series analysis methods are compared with the results of the artificial neural network model which has the best performance by employing different architectures, layer numbers, cell numbers in layers, activation functions and learning methods using the data which is provided from Borsa Istanbul Inc. and covering the dates between 04.02.2005 and 31.12.2015.According to the results of analysis, ARMA (4,4) model performed better than RBF-1-BL artificial neural network model and ARCH (1) model for TL/Dollar contract series. For the Bist-30 contract series, TDNN-1-B-L artificial neural network model has higher predictive performance than ARMA (4.5) and ARCH (1) models.
Derivative markets developed for eliminating uncertainty and risk arising from financial markets can make predictions about the future by using past price movements in case the market is not effective. In this context, in this study, firstly, the effectiveness of the Turkish Derivatives Market was tested by applying the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP) and Kwiatkowski et al. (KPSS) linear unit root tests and Kapetanios et al. (KSS) nonlinear unit root test. As a result of all unit root tests, it was concluded that the series did not show random walk, so that the market was not effective. Then, the method that shows the highest performance is tried to be determined when forecasting the end of day settlement price of the TL/Dollar and Bist-30 contracts which is traded in the Derivatives Market. For this purpose, the forecasting results produced by the time series analysis methods are compared with the results of the artificial neural network model which has the best performance by employing different architectures, layer numbers, cell numbers in layers, activation functions and learning methods using the data which is provided from Borsa Istanbul Inc. and covering the dates between 04.02.2005 and 31.12.2015.According to the results of analysis, ARMA (4,4) model performed better than RBF-1-BL artificial neural network model and ARCH (1) model for TL/Dollar contract series. For the Bist-30 contract series, TDNN-1-B-L artificial neural network model has higher predictive performance than ARMA (4.5) and ARCH (1) models.
Alan : Sosyal, Beşeri ve İdari Bilimler
Dergi Türü : Uluslararası
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|