Published in Scientific Papers. Series "Management, Economic Engineering in Agriculture and rural development", Vol. 21 ISSUE 3
Written by Mehmet Arif ŞAHİNLİ

Using GARCH(p,q) models, in this study our aim is to examine and search the characteristics of volatility of Turkish sugar price. Due to the ARCH effects on price, ARCH(q), GARCH(p,q) and EGARCH(p,q) including these effects on mean and variance equations were estimated. Normal, t-Student, and generalized error distributions with Maximum Likelihood Estimation Method were estimated for these models. Determining the optimal parameters, Marquardt’s algorithm (1963) was used for maximizing the log-likelihood function. Mean absolute percentage error (MAPE), root mean square error (RMSE) and mean absolute deviation (MAD) were used to determine the fit model for making predicting. In this study, we found the best model as a GARCH (1,1) model.

[Read full article] [Citation]

SPUPOPUPNO1