摘要
“碳达峰、碳中和”目标对于煤炭行业“碳减排”提出了巨大挑战,煤炭行业将面临全方位的深度调整。煤炭价格的预测对于有序降低碳排放至关重要,但其波动性和不稳定性受多种影响因素共同作用,因此对煤炭价格的变化做出准确预测较为困难。本文基于深度学习方法,构建多特征下的网络融合模型(CNN-BiLSTM-ARIMA),并应用于秦皇岛动力煤(大同优混5800)平仓价预测。选取判定系数R2、平均绝对百分比误差MAPE、均方根误差RMSE作为评价模型的指标,并与单一长短期记忆网络(LSTM)、双向长短期记忆网络(BiLSTM)及中间网络进行比较分析,实验结果表明本文提出的融合模型具有更好的预测性能和实用性。
The goal of “carbon peak and carbon neutrality” poses a great challenge to the “carbon emission reduction” of the coal industry, and the coal industry will face an all-round deep adjustment. The prediction of coal price is very important for the orderly reduction of carbon emissions, but its volatility and instability are affected by many factors, so it is difficult to make an accurate prediction of the change of coal price. In this paper, based on deep learning method, a multi-feature network fusion model (CNN-BiLSTM-ARIMA) is constructed and applied to the closing price prediction of Qinhuangdao thermal coal (Datong Youming 5800). Coefficient R2, mean absolute percentage error MAPE and root mean square error RMSE were selected as the indexes to eval-uate the model, and compared with single long short-term memory network (LSTM), bidirectional long short-term memory network (BiLSTM) and intermediate network. The experimental results show that the fusion model proposed in this paper has better prediction performance and practicality.
出处
《运筹与模糊学》
2023年第6期7768-7780,共13页
Operations Research and Fuzziology
