Optimal tests in AR(m) time series model
Applications
2013-08-28 v1
Abstract
A method for an evaluation of the error between an unknown parameter and its estimator is developed. Its application enables us to preserve the asymptotic power of a constructed test. Testing problems in AR(1) and ARCH models are studied with a derivation of the asymptotic power function. Also the results are extended to AR(m) time series model.
Cite
@article{arxiv.1308.5767,
title = {Optimal tests in AR(m) time series model},
author = {Tewfik Lounis},
journal= {arXiv preprint arXiv:1308.5767},
year = {2013}
}