Quantum Artificial Intelligence and Oil Price Forecasting: Evidence from the WTI Market

Document Type : Original Article

Authors

1 Associate Professor of Economics, Department of Economics, Faculty of Humanities, Ayatollah Boroujerdi University

2 M.Sc. Student in Economics, Faculty of Humanities, Ayatollah Boroujerdi University

Abstract

This study, within the framework of the Quantum Deep Learning (QDL) approach and utilizing the Quantum Long Short‑Term Memory (QLSTM) model, forecasts the future price of West Texas Intermediate (WTI) crude oil. Given the strategic role of oil in the global economy and the impact of its fluctuations on macroeconomic indicators, developing accurate forecasting models is of great importance for policymakers, investors, and market participants. Daily WTI crude oil futures price data from January 2003 to June 2025 are used. The QLSTM model, which combines deep learning capabilities with quantum computing, is compared with six alternative models: classical time series models (AR, MA), machine learning algorithms (SVR), and traditional deep learning models (LSTM, GRU, MLP). Results show that QLSTM achieves the lowest forecast errors (MSE, RMSE, and Theil’s U) compared to the other models, and the Diebold‑Mariano test statistically confirms its superiority. The findings indicate that QLSTM, by exploiting unique features of quantum computing such as superposition and entanglement, has a high ability to identify complex patterns and extreme volatility in the oil market. These results have important implications for Iranian policymakers in risk management, budgeting, and energy policy design, and also open new avenues for future research.

Keywords

Files

Share

How to cite

Statistics