Manuscript received March 24; accepted May 9, 2025; published June 25, 2025
Abstract—China's carbon price (i.e. carbon emission trading price) is an important part of China's carbon market (i.e. carbon emissions trading market). The study of carbon price can be divided into two categories: the analysis of the influencing factors of carbon price, and the application of the carbon price forecasting models. Studies on influencing factors and carbon price forecasting models are often conducted independently, that is, the application of influencing factors in carbon price forecasting is often neglected. This paper incorporates the key influencing factors into the forecasting model to improve the credibility of the prediction results. Recently, a kind of hybrid forecasting model performs well in many fields and has not yet been used for carbon price prediction. This hybrid model combines ARIMA (Auto-Regressive Integrated Moving Average) with ACNN (Attention-based CNN), LSTM (Long Short-Term Memory), and XGBoost (eXtreme Gradient Boosting), making itself both interpretable and capable of processing extremely complex data. We denote this model as AALX (i.e. ARIMA integrated with ACNN-LSTM and XGBoost) and apply it to the field of carbon price prediction, after taking the key influencing factors of carbon price as the variable of the AALX model. Empirical results indicate that AALX not only retains the existing advantages, but also attains higher credibility.
Keywords—carbon emission allowance, carbon emissions
trading market, Pearson correlation coefficient, ARIMA,
attention mechanism, LSTM, XGBoost
Cite: Yu Hao, Ailing Gu, Chengtuo Lie, and Jiayan Lin, "Forecasting Carbon Price in China Using an ARIMA Integrated with ACNN-LSTM and XGBoost," International Journal of Machine Learning vol. 15, no. 2, pp. 45-50, 2025.
Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).