Serviços Personalizados
Journal
Artigo
Indicadores
-
Citado por SciELO
-
Acessos
Links relacionados
-
Citado por Google
-
Similares em SciELO
-
Similares em Google
Compartilhar
Revista Colombiana de Estadística
versão impressa ISSN 0120-1751
Rev.Colomb.Estad. vol.44 no.2 Bogotá jul./dez. 2021 Epub 01-Set-2021
https://doi.org/10.15446/rce.v44n2.91356
Original articles of research
Forecasting with Multivariate Threshold Autoregressive Models
Pronósticos con modelos multivariados autorregresivos de umbrales
1 Statistics Department, Faculty of Science, Universidad Nacional de Colombia city, Colombia
An important stage in the analysis of time series is forecasting of the interest variables. However, the forecasting in non-linear time series models is not straightforward as in linear time series models because an exact analytical expression for the conditional expectation it is not easy to obtain. In this paper, a procedure for forecasting with multivariate threshold autoregressive(MTAR) models is proposed via the so-called predictive distributions in the Bayesian approach. This strategy gives us the forecasts for the response and exogenous variable vectors. The coverage percentages of the forecast intervals and the variability of the predictive distributions are analyzed in this work. An application in the Hydrology field is presented.
Key words: Bayesian approach; Forecasting; Predictive distributions; Coverage percentages; Multivariate threshold autoregressive Model
Una etapa importante en el análisis de series de tiempo es el pronóstico de las variables de interés. Sin embargo, el pronóstico en modelos de series de tiempo no lineales no es directo como en el caso de modelos lineales de series de tiempo porque obtener la forma analítica exacta de la esperanza condicional no es fácil. En este artículo, un procedimiento de pronóstico con modelos multivariados autorregresivos de umbrales(MTAR) es propuesta vía las llamadas distribuciones predictivas en el enfoque Bayesiano. Esta estrategia nos entrega tanto los pronósticos del vector de respuesta, como el de las variables exógenas. Los porcentajes de cobertura de los intervalos de pronósticos y la variabilidad de las distribuciones predictivas son analizadas en este trabajo. Una aplicación al campo de la hidrología es presentada.
Palabras clave: Enfoque bayesiano; Pronóstico; Distribuciones predictivas; Porcentajes de cobertura; Modelo multivariado Autoregresivo de umbrales
References
Calderon, S. & Nieto, F. (2017), ‘Bayesian analysis of Multivariate Threshold Autoregressive Models with Missing Data’, Communications in Statistics: Theory and Methods 46(1), 296-318. [ Links ]
De Gooijer, J. G. & Vidiella-i Anguera, A. (2004), ‘Forecasting threshold cointegrated systems’, International Journal of Forecasting 20(2), 237-253. [ Links ]
Dellaportas, P., Forster, J. & Ntzoufras, I. (2002), ‘On Bayesian model and variable selection using MCMC’, Statistics and Computing 12(1), 27-36. [ Links ]
Geweke, J. & Amisano, G. (2010), ‘Comparing and evaluating Bayesian predictive distributions of asset returns’, International Journal of Forecasting 26(2), 216-230. [ Links ]
Harvey, A. (1989), Forecasting, structural time series model and the Kalman filter, Cambridge University Press, Cambridge. [ Links ]
Kock, A. B. & Teräsvirta, T. (2011), Forecasting With Nonlinear Time Series Models, in M. P. Clements & D. F. Hendry, eds, ‘The Oxford Handbook of Economic Forecasting’, Oxford University Press, Oxford, chapter 2, pp. 61-87. [ Links ]
Kwon, Y. (2003), Bayesian Analysis of Threshold Autoregressive Models, PhD thesis, University of Tennessee - Knoxville. [ Links ]
Kwon, Y., Bozdogan, H. & Bensmail, H. (2009), ‘Performance of Model Selection Criteria in Bayesian Threshold VAR (TVAR) Models’, Econometric Reviews 28(1-3), 83-101. [ Links ]
Meyn, S. & Tweedie, R. (2009), Markov Chains and Stochastic Stability, Oxford Statistical Science, Cambridge University Press, Cambridge. [ Links ]
Nieto, F. (2005), ‘Modeling Bivariate Threshold Autoregressive Processes in the Presence of Missing Data’, Communications in Statistics - Theory and Methods. 34(4), 905-930. [ Links ]
Nieto, F. (2008), ‘Forecasting with univariate TAR models’, Statistical Methodology 5(3), 263-273. [ Links ]
Tong, H. (2015), ‘Threshold models in time series analysis-Some reflections’, Journal of Econometrics 189(3), 485-491. [ Links ]
Tsay, R. S. (1998), ‘Testing and Modeling Multivariate Threshold Models’, Journal of the American Statistical Association 93(443), 467-482. [ Links ]
Tsay, R,S . & Chen, R. (2018), Nonlinear Time Series Analysis, Wiley Series in Probability and Statistics, John Wiley and Sons. [ Links ]
Vargas, L. (2012), Cálculo de la distribución predictiva en un modelo TAR, Master’s thesis, Universidad Nacional de Colombia, Colombia. [ Links ]
Wu, C. & Lee, J. (2011), ‘Forecasting Time-Varying Covariance with a Robust Bayesian Threshold Model’, Journal of Forecasting 30(5), 451-468. [ Links ]
Received: November 2020; Accepted: May 2021