|
|
 |
| |
|
|
|
Machine Learning Techniques in a Hybrid Forecasting Model for Oil Prices Combining ARIMA |
|
|
|
PP: 1241-1252 |
|
|
|
doi:10.18576/amis/190601
|
|
|
|
Author(s) |
|
|
|
Imad Y. Hamid,
Sulima M. Awad Yousif,
Saud Aljaloud,
Abdelgalal O. I. Abaker,
Halla Z. S. Elemam,
Khalil A. Alruwaitee,
Manahill I. A. Anja,
|
|
|
|
Abstract |
|
|
| Accurately forecasting world oil prices is essential for formulating effective economic strategies and informed energy policies. While traditional time-series models such as the Autoregressive Integrated Moving Average (ARIMA) are widely used for capturing linear patterns in historical data, they often struggle with the nonlinear and complex dynamics characteristic of oil price movements. Conversely, Decision Tree models effectively detect nonlinear relationships but may lack the ability to model long-term dependencies. This paper proposes a hybrid forecasting model that combines ARIMA and Decision Tree techniques to leverage the strengths of both approaches. The ARIMA model is first employed to capture linear trends and produce baseline forecasts, which are then refined by Decision Tree algorithms to address residual nonlinearities. Experimental evaluations conducted on historical oil price data demonstrate that the hybrid ARIMA2013 Decision Tree model consistently outperforms its individual components in predictive accuracy. Model performance is assessed using statistical measures including Mean Absolute Percentage Error (MAPE) and Symmetric Mean Absolute Percentage Error (sMAPE). The results highlight the robustness and effectiveness of the proposed hybrid method in modeling the intricate behavior of global oil prices. This approach provides a reliable and comprehensive decision-support tool for policymakers and industry stakeholders. Future work may extend this framework to other domains where accurate time-series forecasting is critical. |
|
|
|
|
|
|
|
