Harnessing Big Data for Price Predictions in Oil and Gas Markets

This study aims to construct a robust forecasting model for predicting future prices of oil and gas resources, crucialfor diverse industries. Beginning with an exploration of foundational theories pertinent to oil and gas price forecasting, the research focuses on two pivotal forecast theories: the Target Capacity Utilization Principle (TCU) and the Theory of Exhaustible Resources. Employing a recursive simulation model grounded in the Target Capacity Utilization Rule and leveraging historical data spanning from 1987 to 2017, the study forecasts oil and gas prices from 1991 to 2017. The devised methodology results in a Mean Absolute Deviation (MAD) of 12.676, Mean Squared Error (MSE) of 280.92, Mean Absolute Percentage Error (MAPE) of 0.2597, and Mean Percentage Error (MPE) of 0.028. Furthermore, the analysis is extended to monthly forecasts, yielding a MAD of 5.66, MSE of 82.1163, MAPE of 0.1246, and MPE of 0.038. These outcomes underscore the model's effectiveness, especially at the monthly forecasting level.