Non Linear ARX-Models : Probabilistic Properties and Consistent Recursive Estimation

Consider the general ARX(k, q) nonlinear process defined by the recurrence relation Yn = f(Y_(n-1),... y_(n-k), x_n,..., x_(n-q+1)) +([ xi ]_n) where { x_n},{ [xi]_n} are independent i.i.d. sequences. We study some probabilistic properties of this process in the ergodic situation and propose a recursive estimator of the stochastic gradient kind of the function f which is strongly consistent.

MoreLess

Year of publication:	1992-10-01
Authors:	DOUKHAN, P. ; TSYBAKOV, A.
Institutions:	Center for Operations Research and Econometrics (CORE), École des Sciences Économiques de Louvain

More details

Extent:	application/pdf
Series:	CORE Discussion Papers.
Type of publication:	Book / Working Paper
Notes:	The text is part of a series UNIVERSITE CATHOLIQUE DE LOUVAIN, Center for Operations Research and Econometrics (CORE) Number 1992056
Classification:	G - Financial Economics
Source:	RePEc - Research Papers in Economics

Persistent link: https://ebvufind01.dmz1.zbw.eu/10005065354