Statistical Modelling 4 (2004), 19–38

Periodic Markov switching autoregressive models for Bayesian analysis and forecasting of air pollution

Luigi Spezia
Department of Statistics
Athens University of Economics and Business
Greece

Roberta Paroli
Istituto di Statistica
Università Cattolica S.C. di Milano
Italy
eMail: roberta.paroli@unicatt.it

Petros Dellaportas
Department of Statistics
Athens University of Economics and Business
Greece

Abstract:

Markov switching autoregressive models (MSARMs) are eficient tools to analyse linear and non-Gaussian time series. A special MSARM with two harmonic components is proposed to analyse periodoc time series. We present a full Bayesian analysis based on Gibbs sampling algorithm for model choice and the estimations of the unknown parameters, missing data and predictive distributions. The implementation and modelling steps are developed by tackling the problem of the hidden states labeling by means of random permutation sampling and constrained permutation sampling. We apply MSARMs to study a data set about air pollution which presents periodicities since the hourly mean concentration of carbon monoxide varies according to the dynamics of the 24 day-hours and of the year. Hence, we introduce in the model both a hidden state-dependent daily component and a state-independent yearly component, giving rise to periodic MSARMs.

Keywords:

Hidden Markov chain; harmonic components; Gibbs sampling; label switching; permutation sampling; carbon monoxide.
 

Downloads:

Data and software in zipped archive


back