Browse Journals

Journal Menu

EFFICIENT ESTIMATION IN RESTRICTIVE PERIODIC EXPAR(1) MODELS 

Author(s) : H. Dridi, M. Merzougui and A. Chadli
Pages : [45] - [74]
Received : February 4, 2017; Revised May 10, 2017
Communicated by : Professor Marcelo Bourguignon

Abstract

This paper is devoted to study the problem of estimation in the restricted periodic exponential autoregressive (PEXPAR(1)) model. The asymptotic optimality of the procedure is shown via local asymptotic normality (LAN). Once the LAN property is proved, we construct a parametric locally asymptotically minimax LAM estimator. Using these results, we construct the adaptive estimators for the parameters when the innovation density is unknown. The performance of the established estimators is shown via small simulation 

Keywords

periodically correlated process, restricted periodic exponential autoregressive model, local asymptotic normality (LAN), local asymptotic minimax (LAM) estimator, adaptive estimation.