Estimation of Parametric Models with Conditional Heteroscedastic Errors
Résumé
We consider a model with conditional heteroscedastic errors. The model requires only the form of conditional mean and conditional variance functions to be specified. We propose an effective approach for fitting this class of model. Our estimator is deduced from quasi-likelihood concept using an iterative and adaptive procedure. The convergence properties are establishe- d. Finally, our method and widely used estimators are compared via numerical experiments.