A poisson state space model for modeling football matches between Brazil and Argentina.

Authors

Keywords:

Non-Gaussian mode, Bayesian and classical inference, exact marginal likelihood, soccer matches

Abstract

This article presents a Poisson state space model for modeling the historical soccer
matches between Brazil and Argentina. This model allows the exact calculation of marginal likelihood function in a easy fashion as well as the predictive, smoothing and filtering distributions
for the latent variable. It is not necessary to use approximations, which is very common in the
literature under the non-Gaussian context. The insertion of covariates and treatment of data
irregularity may be done in a natural way without problems. A covariate called “Factor field”,
that indicates the match place (home or outside), is inserted into the model and influences the
outcome of the match. When the team plays at home, it has more chance to score goals, which
was already expected in general. The results are very satisfactory and illustrate well the proposed
model.

Author Biography

Thiago Rezende dos Santos, UFMG

Depto Estatística/UFMG

References

FARIAS, F. F. (2008). Análise e Previsão de Resultados de Partidas de Futebol. Departamento de Métodos Estatísticos, Universidade Federal do Rio de Janeiro, 2008.

GAMERMAN, D. & LOPES, H. F. (2006). Markov Chain Monte Carlo: Stochastic Simulation for Bayesian Inference. (2nd edition). London: Chapman and Hall.

GAMERMAN, D., Santos, T. R., FRANCO, G.C. (2013). A non-Gaussian family of state-space models with exact marginal likelihood. Technical Report, RTP- 01/2013, Departamento de Estatística, Universidade Federal de Minas Gerais. Available at http://www.est.ufmg.br/portal/arquivos/rts

HARVEY, A.C. (1989). Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press.

RUE, H e SALVESEN, O. (2000). Prediction and retrospective analysis of soccer matches in a league. Norwegian University of Science and Technology. Trondheim, Noruega.

SOUZA JR, O. G. e GAMERMAN, D. (2004). Previsão de partidas de futebol usando modelos dinâmicos. Anais do XXXVI SBPO. São João del Rey - MG.

WEST, M. e HARRISON, J. (1997). Bayesian Forecasting and Dynamic Models. New York: Springer.

Published

31-12-2013

How to Cite

Santos, T. R. dos. (2013). A poisson state space model for modeling football matches between Brazil and Argentina. Sigmae, 2(1), 42–47. Retrieved from https://publicacoes.unifal-mg.edu.br/revistas/index.php/sigmae/article/view/109

Issue

Section

Applied Statistics