Combined Forecasts Using the Akaike Weights

Mariola Piłatowska

DOI: http://dx.doi.org/10.12775/DEM.2009.001

Abstract


The focus in the paper is on the information criteria approach and especially the Akaike information criterion which is used to obtain the Akaike weights. This approach enables to receive not one best model, but several plausible models for which the ranking can be built using the Akaike weights. This set of candidate models is the basis of calculating individual forecasts, and then for combining forecasts using the Akaike weights. The procedure of obtaining the combined forecasts using the AIC weights is proposed. The performance of combining forecasts with the AIC weights and equal weights with regard to individual forecasts obtained from models selected by the AIC criterion and the a posteriori selection method is compared in simulation experiment. The conditions when the Akaike weights are worth to use in combining forecasts were indicated. The use of the information criteria approach to obtain combined forecasts as an alternative to formal hypothesis testing was recommended.


Keywords


combining forecasts, weighting schemes, information criteria

Full Text:

PDF

References


Akaike, H. (1973), Information Theory as an Extension of the Maximum Likelihood Principle, [in:] Petrov, B. N., Csaki, F., Second International Symposium on Information Theory, Akademia Kiado, Budapest.

Akaike, H. (1978), On the Likelihood of a Time Series Model, The Statistician, 27, 217–235. DOI: http://dx.doi.org/10.2307/2988185

Armstrong, J. S. (2001), Principles of Forecasting, Springer.

Atkinson, A. C. (1980), A Note on the Generalized Information Criteria for Choice of a Model, Biometrika, 67 (2), 413–418.

Bates, J. M., Granger, C. W. J. (1969), The Combinations of Forecasts, Operations Research Quarterly, 20, 415–468. DOI: http://dx.doi.org/10.1057/jors.1969.103

Burnham, K. P., Anderson, D. R. (2002), Model Selection and Multimodel Inference, Springer.

Burnham, K. P., Anderson, D. R. (2004), Multimodel Inference. Understanding AIC and BIC in Model Selection, Sociological Methods and Research, vol. 33 (2), 261–304.

Jagannathan, R. Ma, T. (2003), Risk Reduction in Large Portfolios: Why Imposing the Wrong Constraints Helps, The Journal of Finance, 58 (4), 1651–1684. DOI: http://dx.doi.org/10.1111/1540-6261.00580

Kapetanios, G., Labhard, V., Price, S. (2008), Forecasting using Bayesian and Informationtheoretic Model Averaging: an Application to U.K. Inflation, Journal of Business and Economics Statistics, 26 (1), 33–41.

Kitchen, J., Monaco, R. (2003), Real-Time Forecasting in Practice, Business Economics, 38 (4), 10–19.

Marcellino, M. (2004), Forecast Pooling for Short Time Series of Macroeconomic Variables, Oxford Bulletin of Economic and Statistics, 66, 91–112.

Min, C. K., Zellner, A. (1993), Bayesian and Non-Bayesian Methods for Combining Models and Forecasts with Applications to Forecasting International Growth Rates, Journal of Econometrics, 53 (1–2), 89–118.

Stock, J. H., Watson, M. (2004), Combination Forecasts of Output Growth in a Seven-Country Data Set, Journal of Forecasting, 8, 230–251.

Stock, J. H., Watson, M. (2006), Forecasting with Many Predictors, [in:] Elliott, G., Granger, C. W. J., Timmermann, A. (ed.), Handbook of Economic Forecasting, Elsevier.

Swanson, N. R., Zeng, T. (2001), Choosing Among Competing Econometric Forecasts: Regression-based Forecast Combination using Model Selection, Journal of Forecasting, 20, 425–440.

Timmermann, A. (2006), Forecast Combinations, [in:] Elliott G., Granger C. W. J., Timmermann A. (ed.), Handbook of Economic Forecasting, ch. 4, Elsevier.






ISSN (print) 1234-3862
ISSN (online) 2450-7067

Partnerzy platformy czasopism