Crude oil price forecasting based on hybridizing wavelet multiple linear regression model, particle swarm...

Area of Science:

• Economics
• Financial Modeling
• Time Series Analysis

Background:

• Crude oil prices significantly influence the global economy, impacting financial markets and risk assessment.
• Accurate forecasting of crude oil prices is crucial for economic stability and investment strategies.
• Existing models often struggle to capture the complex dynamics of oil price fluctuations.

Purpose of the Study:

• To propose a novel hybrid model for enhanced crude oil price forecasting.
• To integrate wavelet transform, principal component analysis, and particle swarm optimization for improved prediction accuracy.
• To evaluate the proposed model's performance against established forecasting techniques.

Main Methods:

• A hybrid model combining Mallat wavelet transform for time series decomposition and multiple linear regression (MLR) for forecasting.
• Principal Component Analysis (PCA) applied to subseries data for dimensionality reduction.
• Particle Swarm Optimization (PSO) utilized for optimizing MLR model parameters.
• Case study using daily West Texas Intermediate (WTI) crude oil prices.

Main Results:

• The proposed Wavelet-Multiple Linear Regression (WMLR) model demonstrated superior performance in crude oil price forecasting.
• Comparative analysis showed WMLR outperforming individual MLR, ARIMA, and GARCH models.
• Statistical measures confirmed the enhanced predictive capability of the hybrid approach.

Conclusions:

• The hybrid WMLR model offers a robust and effective solution for crude oil price forecasting.
• Integrating wavelet decomposition and optimization techniques significantly improves prediction accuracy.
• This approach provides valuable insights for economic and financial decision-making in the energy sector.