Optimal Estimation of Multivariate ARMA Models
- Martha White, University of Alberta
- Junfeng Wen
- Michael Bowling, University of Alberta
- Dale Schuurmans, AICML
Autoregressive moving average (ARMA) models are a fundamental tool in timeseries analysis that offer intuitive modeling capability and efficient predictors. Unfortunately, the lack of globally optimal parameter estimation strategies for these models remains a problem:application studies often adopt the simpler autoregressive model that can be easily estimated by maximizing (a posteriori) likelihood. We develop a (regularized, imputed) maximum likelihood criterion that admits efficient global estimation via structured matrix norm optimization methods. An empirical evaluation demonstrates the benefits of globally optimal parameter estimation over local and moment matching approaches.
Citation
M. White, J. Wen, M. Bowling, D. Schuurmans. "Optimal Estimation of Multivariate ARMA Models". National Conference on Artificial Intelligence (AAAI), (ed: Blai Bonet, Sven Koenig), pp 3080-3086, January 2015.| Keywords: | |
| Category: | In Conference |
| Web Links: | AAAI |
BibTeX
@incollection{White+al:AAAI15,
author = {Martha White and Junfeng Wen and Michael Bowling and Dale
Schuurmans},
title = {Optimal Estimation of Multivariate ARMA Models},
Editor = {Blai Bonet, Sven Koenig},
Pages = {3080-3086},
booktitle = {National Conference on Artificial Intelligence (AAAI)},
year = 2015,
}Last Updated: February 14, 2020Submitted by Sabina P
