Advanced Arimax Modeling For Opec Oil Forecasting Using Multivariate Wavelet Techniques
DOI:
https://doi.org/10.64252/wz878a51Keywords:
ARIMAX modeling, OPEC oil forecasting, multivariate wavelet techniques, time series analysis, wavelet transformation, forecasting accuracy, model order optimization, multicollinearity testing and time series simulation.Abstract
This study develops a robust oil price forecasting framework by integrating wavelet transformation, multicollinearity diagnostics, and ARIMAX modeling enhanced with Monte Carlo simulation. The dataset includes key economic indicators such as OPEC production, global demand and supply, GDP figures, and oil transportation costs. Preliminary analysis revealed strong multicollinearity among explanatory variables, which was successfully mitigated using Haar wavelet decomposition. Stationarity of the oil price series was confirmed through the Augmented Dickey-Fuller (ADF) test after first differencing. Several ARIMAX model configurations were tested, with ARIMA(2,1,1) emerging as the optimal model based on AIC, RMSE, and MAPE criteria. Monte Carlo simulations, conducted over 1,000 iterations, demonstrated the model's forecasting stability and predictive reliability. Forecasts for the 2025–2026 period suggest a relatively stable oil market, with price projections ranging between $76.96 and $79.57 per barrel. The study's methodological framework offers a valuable approach for short-term energy market forecasting and supports informed decision-making by policymakers and stakeholders in the oil industry.
