摘要
碳交易价格波动频繁,非线性随机性强、具有动态变化特征,准确预测碳交易价格对于具有碳交易活动需求的企业而言是极其必要的。因此选取深圳碳交易价格和影响碳交易价格的十个因素,构建LASSO-BP神经网络模型对深圳碳交易价格进行预测研究,并与传统ARIMA模型进行分析对比。结果表明,LASSO-BP神经网络模型可以在短期精准预测深圳碳交易价格,LASSO-BP神经网络模型相比于ARIMA模型有更高的R2和更低的MSE,能够更有效地提高碳交易价格的预测精度和预测稳定性,并且交易价格呈现下降的趋势,这也与我国一直以来倡导节能减排、绿色金融的发展方式不谋而合。
The carbon trading price fluctuates frequently,exhibits strong nonlinear randomness,and undergoes dynamic changes.Therefore,it is crucial for enterprises engaged in carbon trading to accurately predict these prices.This study selects the Shenzhen carbon trading price and ten influencing factors,constructs a LASSO-BP neural network model for forecasting,and compares its performance with the traditional ARIMA model.The results indicate that the LASSO-BP neural network model can accurately predict Shenzhen’s carbon trading price in the short term.Compared to the ARIMA model,the LASSO-BP neural network model exhibits a higher R2 and a lower MSE,enhancing both prediction accuracy and forecast stability.Additionally,the carbon trading price shows a downward trend.This aligns with China's ongoing advocacy for energy conservation,emission reduction,and green finance.
作者
夏正兰
王昭媛
XIA Zhenglan;WANG Zhaoyuan(School of Science,Civil Aviation Flight University of China,Chengdu 641419,China;Xinjiang Institute of Science and Technology,Korla 841000,Xinjiang China)
出处
《河南科学》
2025年第3期427-436,共10页
Henan Science
基金
民航飞行技术与飞行安全重点实验室项目(FZ2022ZZ05,FZ2022ZX35,FZ2022ZX60)
中央高校基本科研业务费资助项目(PHD2023-054,PHD2023-056)
新疆科技学院校级科研项目(2024-KYPT28)。
