MODELING AND FORECASTING EXPORTS FROM BRAZIL MAKING USE OF TIME SERIES

Main Article Content

Claudio Javier TABLADA
Fernando Arturo Peña RAMÍREZ
Giannini Italino Alves VIEIRA
Josimar M. de VASCONCELOS

Abstract

Using available tools for time series and with the support of statistical software R, we intend to model and make predictions for h periods ahead (h  = 1; 3; 6) under export data of Brazil from January 1995 to November 2012. For this end, dierent modeling techniques were used, as exponential smoothing algorithms and stochastic models. Initially, it was considered to be the Holt-Winters algorithm in its dierent versions and, additionally, models of  type SARIMA and SARIMAX were selected. Finally, the power of forecast of the dierent models is analyzed using the relative errors of forecast as measure of evaluation.

Article Details

How to Cite
TABLADA, C. J., RAMÍREZ, F. A. P., VIEIRA, G. I. A., & VASCONCELOS, J. M. de. (2016). MODELING AND FORECASTING EXPORTS FROM BRAZIL MAKING USE OF TIME SERIES. Brazilian Journal of Biometrics, 34(1), 33–48. Retrieved from https://biometria.ufla.br/index.php/BBJ/article/view/89
Section
Articles