摘要
在“双碳”背景下,煤炭价格受到多种非线性和非平稳因素的影响,煤炭价格的准确可靠预测对增强煤炭市场的稳定性愈发重要。以LSTM神经网络为基础算法研究环渤海动力煤价格指数波动状态,并实现对该指数未来趋势的准确预测;采用混合核密度估计方法探讨煤炭价格指数未来趋势的不确定性,实现了对该价格指数未来价格波动区间分布的概率评估,可为企业决策人员提供全面、高精度的预测信息。经过与经典的BPNN、RNN模型的对比,新模型使环渤海动力煤价格指数预测值的准确度提高,同时能准确提供企业未来价格波动的概率评估结果,为其经营实践提供决策支持。
In the context of the"dual-carbon"initiative,coal prices are influenced by various nonlinear and non-stationary factors.Accurate and reliable prediction of coal prices becomes increasingly important for enhancing the stability of the coal market.This study utilizes the Long Short-Term Memory(LSTM)neural network as the foundational algorithm to investigate the fluctuation patterns of the Circum-Bohai-Sea thermal coal price index and achieve accurate forecasting of its future trends.The research employs a hybrid kernel density estimation method to explore the uncertainty of future trends in the coal price index.This approach facilitates a probability assessment of the future price fluctuation range distribution for the index,providing comprehensive and high-precision predictive information for decision-makers in enterprises.In comparison with classical BPNN and RNN models,the new model enhances the accuracy of predictions for the Circum-Bohai-Sea thermal coal price index.Simultaneously,it accurately provides probability assessments of future price fluctuations for enterprises,offering decision support for their operational practices.
出处
《价格理论与实践》
北大核心
2024年第2期42-46,125,共6页
Price:Theory & Practice
基金
企业精益化管理应用系统开发(H2022-172)
中国华能集团有限公司基础管理能力提升咨询服务(H2021-176)
关键词
环渤海动力煤价格预测
LSTM模型
核密度估计
波动性预测
Bohai Rim thermal coal price prediction
LSTM model
kernel density estimation
volatility pvrediction
