A sentiment-driven three-stage approach for multi-scale carbon price prediction
Abstract An accurate calculation method of carbon trading price is of great significance to strengthening energy saving and emission reduction. Due to the nonlinear and non-stationary characteristics of the carbon price, it is difficult to predict the carbon price accurately. This paper proposes a n...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-06-01
|
| Series: | Discover Sustainability |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s43621-025-01258-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract An accurate calculation method of carbon trading price is of great significance to strengthening energy saving and emission reduction. Due to the nonlinear and non-stationary characteristics of the carbon price, it is difficult to predict the carbon price accurately. This paper proposes a new hybrid model for carbon trading price forecasting. The model fuses complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) with extreme gradient boosting (XGBoost) and long short-term memory (LSTM) networks, and leverages SnowNLP to derive sentiment scores from news text and the Baidu Index. To demonstrate the superiority of the proposed model, 5 chinese carbon emissions trading markets are selected for the predictions. The model shows better performance across all markets, improving by 4.20% to 17.89% over the CEEMDAN-LSTM model and outperforming other benchmarks. Furthermore, ablation experiments and parametric sensitivity analyses were carried out to verify the contribution of each component and the overall model’ s robustness. It offers a reliable and stable forecasting tool for stakeholders in the carbon market. |
|---|---|
| ISSN: | 2662-9984 |