Abstract:
In this study we apply the Generalized Autoregressive Conditional Heteroscedas-tic (GARCH) process to obtain the homoscedasticity among resideuals obtained from asymmetric moving average time series models given in Mahmud (1989). We use the fitted residuals depend on the previous residuals. Using GRACH process, the heteroscedasticity among residuals has been reduced to homoscedasticity. Also the value of the log likelihood function is increased. Finally after fitting asymmetric time series models with GARCH errors on the real time series, the ACF of the fitted squared residuals were found insignificant.
Page(s):
111-119
DOI:
DOI not available
Published:
Journal: Karachi University Journal of Science, Volume: 19, Issue: 1--2, Year: 1991