Il giorno lunedì 2 maggio alle ore 16.00 presso la Aula Seminari del DISMEQ, al quarto piano dell'edificio U7, il prof. Yury Kutoyants della Universitè du Maine, Le Mans, terrà un seminario su

"Threshold autoregressive time series"

Abstract
We present a review of the several problems of parameter estimation for the threshold nonlinear autoregressive time series. The main problem is the threshold estimation. This problem of parameter estimation is singular because the likelihood ratio function is discontinuous, the Fisher information is infinite and the rate of convergence is of order n. We describe the asymptotic behavior of the maximum likelihood estimators (MLE) and Bayesian estimators (BE). It is shown that the BE are asymptotically efficient. The results of numerical simulations are presented. Some generalizations on continuous-time threshold models are discussed.

Tutti gli interessati sono invitati a partecipare.

Prof. Fabio Bellini
Department of Statistics and Quantitative Methods
University of Milano-Bicocca
Via Bicocca degli Arcimboldi 8, 20126 Milano
0039-2-64483119
http://www.economia.unimib.it/bellini
http://scholar.google.it/citations?user=P61L8P4AAAAJ&hl=it