Autoregressive Model of Adaptive Integration

• Arthur Ronald CEFET-RJ
• Rebecca Salles CEFET-RJ
• Kele Belloze CEFET-RJ
• Dayse Pastore CEFET-RJ
• Eduardo Ogasawara CEFET-RJ

DOI: https://doi.org/10.5753/sbbd.2019.8819

Abstract

Several preprocessing techniques combined with time series models have been used to predict non-stationary time series. The study of the mathematical and statistical properties of the data and the preprocessing techniques can help in the adjustment of machine learning models. Such a study, however, may not be easily obtained. Linear models enable the interpretation of such properties. This article introduces and analyzes, based on proof of concept, a new linear model applied to stationary time series that are built by using adaptive normalization. The model allows the use of autoregressive models with sliding windows of data that preserve the properties of the original series and allow the observation of its inertia. The model was able to present superior prediction performance to other linear models consolidated in the literature, specially in short-term horizons.

Keywords: Autoregressive Model, Adaptive Integration

References

Akaike, H. (1992). Information Theory and an Extension of the Maximum Likelihood Principle. In Kotz, S. and Johnson, N. L., editors, Breakthroughs in Statistics: Foundations and BasicTheory, Springer Series in Statistics, pages 610–624. Springer New York, New York, NY. DOI: https://doi.org/10.1007/978-1-4612-0919-5_38

Box, G. E. P., Jenkins, G. M., and Reinsel, G. C. (2008). Time Series Analysis: Forecasting and Control. Wiley, Hoboken, N.J, 4 edition. DOI: https://doi.org/10.1002/9781118619193

Brockwell, P. J. and Davis, R. A. (2016). Introduction to Time Series and Forecasting. Springer International Publishing, Cham, 3 edition. DOI: https://doi.org/10.1007/978-3-319-29854-2

Enders, W. (2015). Applied Econometric Time Series. Wiley, 4 edition.

Hurvich, C. and Tsai, C. (1989). Regression and time series model selection in small samples. Biometrika, 76(2):297–307. DOI: https://doi.org/10.1093/biomet/76.2.297

Hyndman, R. and Khandakar, Y. (2008). Automatic time series forecasting: The forecast package for R. Journal of Statistical Software, 27(3):1–22. DOI: https://doi.org/10.18637/jss.v027.i03

Ogasawara, E., Martinez, L., De Oliveira, D., Zimbrão, G., Pappa, G., and Mattoso, M. (2010). Adaptive Normalization: A novel data normalization approach for non-stationary time series. In Proceedings of the International Joint Conference on Neural Networks. DOI: https://doi.org/10.1109/IJCNN.2010.5596746

Salles, R., Belloze, K., Porto, F., Gonzalez, P., and Ogasawara, E. (2019). Nonstationary time series transformation methods: An experimental review. Knowledge-Based Systems, 164:274–291. DOI: https://doi.org/10.1016/j.knosys.2018.10.041

Tsay, R. S. (2010). Analysis of Financial Time Series. Wiley, Cambridge, Mass, 3 edition. DOI: https://doi.org/10.1002/9780470644560

Woodward, W. A., Gray, H. L., and Elliott, A. C. (2017). Applied Time Series Analysis with R. CRC Press, Boca Raton, 2 edition. DOI: https://doi.org/10.1201/9781315161143