A New Look at Bayesian Identification of Moving Average Models

Abstract

In this paper we review the existing Bayesian identification techniques for the moving average (MA) models that can be classified as testing based identification and posterior mass function based identification. In order to improve the Bayesian identification of MA models, we present a new Bayesian identification method that is based on the posterior mass function of the MA models order.

Publication
In Proceedings of the 53rd Annual International Conference of Statistics, Computer Science and Operations Research, ISSR, Cairo University, Egypt
comments powered by Disqus