Nonasymptotic bounds for autoregressive time series modeling
Goldenshluger, Alexander ; Zeevi, Assaf
Ann. Statist., Tome 29 (2001) no. 2, p. 417-444 / Harvested from Project Euclid
The subject of this paper is autoregressive (AR) modeling of a stationary, Gaussian discrete time process, based on a finite sequence of observations. The process is assumed to admit an AR($\infty$) representation with exponentially decaying coefficients. We adopt the nonparametric minimax framework and study how well the process can be approximated by a finite­order AR model. A lower bound on the accuracy of AR approximations is derived, and a nonasymptotic upper bound on the accuracy of the regularized least squares estimator is established. It is shown that with a “proper” choice of the model order, this estimator is minimax optimal in order. These considerations lead also to a nonasymptotic upper bound on the mean squared error of the associated one­step predictor. A numerical study compares the common model selection procedures to the minimax optimal order choice.
Publié le : 2001-04-14
Classification:  Autoregressive approximation,  minimax risk,  rates of convergence,  62G05,  62M10,  62M20
@article{1009210547,
     author = {Goldenshluger, Alexander and Zeevi, Assaf},
     title = {Nonasymptotic bounds for autoregressive time series
			 modeling},
     journal = {Ann. Statist.},
     volume = {29},
     number = {2},
     year = {2001},
     pages = { 417-444},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1009210547}
}
Goldenshluger, Alexander; Zeevi, Assaf. Nonasymptotic bounds for autoregressive time series
			 modeling. Ann. Statist., Tome 29 (2001) no. 2, pp.  417-444. http://gdmltest.u-ga.fr/item/1009210547/