Forecasting Wheat Prices Based on Past Behavior: Comparison of Different Modelling Approaches
Being able to accurately forecast the evolution of wheat prices can be a valuable tool. Most of the published works apply classical forecasting models to wheat price time series, and they do not always perform out-of-sample testing. This work compares five modelling approaches for wheat price foreca...
| Autor principal: | |
|---|---|
| Outros Autores: | |
| Formato: | bookPart |
| Idioma: | eng |
| Publicado em: |
2019
|
| Assuntos: | |
| Texto completo: | http://hdl.handle.net/10316/88877 |
| País: | Portugal |
| Oai: | oai:estudogeral.sib.uc.pt:10316/88877 |
| Resumo: | Being able to accurately forecast the evolution of wheat prices can be a valuable tool. Most of the published works apply classical forecasting models to wheat price time series, and they do not always perform out-of-sample testing. This work compares five modelling approaches for wheat price forecasts, using only past values of the time series. The models performance is assessed considering out-of-sample data only, by considering a sliding and growing time window that will define the data used to determine the models parameters, and the data used for out-of-sample forecasts. |
|---|