A New Two-Stage Approach with Boosting and Model Averaging for Interval-Valued Crude Oil Prices Forecasting in Uncertainty Environments

In view of the intrinsic complexity oil market, crude prices are influenced by numerous factors that make forecasting very difficult. Recognizing this challenge, approaches have been introduced, but little work has done concerning interval-valued prices. To capture underlying characteristics price movements, paper proposes a two-stage procedure to forecast time series, which generalizes point-valued forecasts incorporate uncertainty and variability. The empirical results show our proposed approach significantly outperforms all benchmark models in terms both accuracy robustness analysis. These can provide references for decision-makers understand trends improve efficiency economic activities.