نویسندگان
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
This study analyzes and compares a general class of Autoregressive Conditional Heteroscedasticity (G) ARCH models during 126 month in Tehran Stock Exchange index (TEDPIX). The results of models confirmed the clustering volatility, asymmetric relation and nonlinearity property in market returns time series. Then the (G) ARCH models enhanced by artificial neural networks. Results suggest that ANN-PGARCH-M, ANN-EGARCH-M and ANN-GJR-GARCH have the least forecasting errors and the volatility direction comparison demonstrates that hybrid models are more excellent than basic (G) ARCH models. According to the results, ANN-APGARCH, ANN-EGARCH-M and ANN-PGARCH-M provide significant improvement in forecasting.
کلیدواژهها [English]