Article ID: | iaor1994770 |
Country: | United Kingdom |
Volume: | 12 |
Issue: | 6 |
Start Page Number: | 525 |
End Page Number: | 538 |
Publication Date: | Aug 1993 |
Journal: | International Journal of Forecasting |
Authors: | Brasil G.H., Souza R.C. |
Keywords: | Bayesian forecasting |
This article deals with a fully Bayesian approach to describe the cyclical behaviour of a univariate time series. A damped sine wave where both the period and the damping factor are time varying is assumed as the underlying mathematical model for the cyclical component. The model is applied to two real data sets; the annual rainfall observations in Fortaleza, Brazil, and the annual Wolf sunspot numbers.