Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Multi-Step Brent Oil Price Forecasting with Ensemble Multi-Scenario Bi-GRU Networks

Mohammed Alruqimi, Luca Di Persio

TL;DR

This work tackles the challenge of accurate multi-step Brent crude oil price forecasting under volatility by evaluating a range of deep-learning architectures and external predictors, then introducing an ensemble ERS-Bi-GRU model that fuses three Bi-GRU forecasts via sentiment and residual components. The approach demonstrates that sequence-based models often outperform transformer-based ones for time-series forecasting in this domain, with the SENT signal and the residual-enhanced ensemble delivering the strongest three-day ahead performance (e.g., MAE ≈ 1.04, RMSE ≈ 1.41). Extensive experiments on a COVID-era dataset illustrate that the proposed ESR-Bi-GRU surpasses state-of-the-art baselines like Autoformer and TimesNet, highlighting the practical value of sentiment and carefully constructed ensembles for energy-market forecasting. The study provides a framework for integrating external factors, residual analysis, and ensemble techniques to improve robustness in volatile financial time series.

Abstract

Despite numerous research efforts in applying deep learning to time series forecasting, achieving high accuracy in multi-step predictions for volatile time series like crude oil prices remains a significant challenge. Moreover, most existing approaches primarily focus on one-step forecasting, and the performance often varies depending on the dataset and specific case study. In this paper, we introduce an ensemble model to capture Brent oil price volatility and enhance the multi-step prediction. Our methodology employs a two-pronged approach. First, we assess popular deep-learning models and the impact of various external factors on forecasting accuracy. Then, we introduce an ensemble multi-step forecasting model for Brent oil prices. Our approach generates accurate forecasts by employing ensemble techniques across multiple forecasting scenarios using three BI-GRU networks.Extensive experiments were conducted on a dataset encompassing the COVID-19 pandemic period, which had a significant impact on energy markets. The proposed model's performance was evaluated using the standard evaluation metrics of MAE, MSE, and RMSE. The results demonstrate that the proposed model outperforms benchmark and established models.

Enhancing Multi-Step Brent Oil Price Forecasting with Ensemble Multi-Scenario Bi-GRU Networks

TL;DR

This work tackles the challenge of accurate multi-step Brent crude oil price forecasting under volatility by evaluating a range of deep-learning architectures and external predictors, then introducing an ensemble ERS-Bi-GRU model that fuses three Bi-GRU forecasts via sentiment and residual components. The approach demonstrates that sequence-based models often outperform transformer-based ones for time-series forecasting in this domain, with the SENT signal and the residual-enhanced ensemble delivering the strongest three-day ahead performance (e.g., MAE ≈ 1.04, RMSE ≈ 1.41). Extensive experiments on a COVID-era dataset illustrate that the proposed ESR-Bi-GRU surpasses state-of-the-art baselines like Autoformer and TimesNet, highlighting the practical value of sentiment and carefully constructed ensembles for energy-market forecasting. The study provides a framework for integrating external factors, residual analysis, and ensemble techniques to improve robustness in volatile financial time series.

Abstract

Despite numerous research efforts in applying deep learning to time series forecasting, achieving high accuracy in multi-step predictions for volatile time series like crude oil prices remains a significant challenge. Moreover, most existing approaches primarily focus on one-step forecasting, and the performance often varies depending on the dataset and specific case study. In this paper, we introduce an ensemble model to capture Brent oil price volatility and enhance the multi-step prediction. Our methodology employs a two-pronged approach. First, we assess popular deep-learning models and the impact of various external factors on forecasting accuracy. Then, we introduce an ensemble multi-step forecasting model for Brent oil prices. Our approach generates accurate forecasts by employing ensemble techniques across multiple forecasting scenarios using three BI-GRU networks.Extensive experiments were conducted on a dataset encompassing the COVID-19 pandemic period, which had a significant impact on energy markets. The proposed model's performance was evaluated using the standard evaluation metrics of MAE, MSE, and RMSE. The results demonstrate that the proposed model outperforms benchmark and established models.
Paper Structure (19 sections, 3 equations, 6 figures, 5 tables)

This paper contains 19 sections, 3 equations, 6 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Literature review
  3. Preliminaries
  4. Dataset
  5. Dataset description
  6. Correlation-based Feature Selection
  7. Data preparation
  8. Methodology
  9. Experiments
  10. Error evaluation metrics
  11. Experimental design
  12. Findings and Discussion
  13. LSTM and GRU models
  14. Transformers and TimesNet Experiments
  15. External variables
  16. ...and 4 more sections

Figures (6)

  • Figure 1: Brent oil price trend from 2012 to 2021
  • Figure 2: Heatmap Spearman Correlation
  • Figure 3: Methodology flowchart
  • Figure 4: Crude Brent oil price components
  • Figure 5: One-ahead day forecasts using SENT-GRU, SENT-LSTM models
  • ...and 1 more figures