On the maximum likelihood estimation in the case of dependent random variables
Résumé
Our approach, based on a general theorem about asymptotic separation by J. Geffroy, requires neither any regularity assumption of the likelihood function nor any version of the law of large numbers for martingales. Furthermore, it is shown that the rate of convergence of the estimator is exponential.
Domaines
Mathématiques [math]Origine | Accord explicite pour ce dépôt |
---|