| Article ID: | iaor1989844 |
| Country: | Netherlands |
| Volume: | 41 |
| Issue: | 1 |
| Start Page Number: | 15 |
| End Page Number: | 22 |
| Publication Date: | Jul 1989 |
| Journal: | European Journal of Operational Research |
| Authors: | Boas J. |
| Keywords: | statistics: regression |
Making forecasts of the price development of cocoa beans using conventional regression models proved to be unsatisfactory because of the instability of the cocoa market. This paper discusses why the conventional model failed and how this forecasting problem was tackled by applying the Kalman filter technique, a model developed for predictions in non-stationary situations.