基于GWO-LSTM模型的辽宁省工业碳排放预测及影响因素研究

Prediction of Industrial Carbon Peaking in Liaoning Province based on GWO-LSTM Model

科学预测工业碳排放对于低碳发展具有重要意义。基于灰色关联分析法,选择人口、人均工业GDP、城市化率、能源消耗、工业开放程度作为辽宁省工业碳排放影响因素,并通过STIRPAT模型,定量分析各影响因素与碳排放量的关系,在此基础上建立灰狼优化算法优化的长短期记忆神经网络(GWO-LSTM)模型对辽宁省工业碳排放进行预测。研究结果表明:人口,城市化,能源消耗,人均工业GDP,工业开放程度每增加1%,碳排放量将相应增加1.04%,0.81%,0.38%,0.27%,0.18%;GWO-LSTM工业碳排放预测模型的决定系数为0.996 8,高于原始的LSTM预测模型以及SVR预测模型。

Scientific prediction of industrial carbon emissions is of great significance for low-carbon development.Based on the grey correlation analysis method,population,per capita industrial GDP,urbanization rate,energy consumption and industrial openness were selected as the influencing factors of industrial carbon emissions in Liaoning Province,and the relationship between each influencing factor and carbon emissions was quantitatively analyzed through the STIRPAT model.On this basis,a long short-term memory neural network(GWO-LSTM)model optimized by the gray wolf optimization algorithm was established to predict the industrial carbon emissions of Liaoning Province.The results show that for every 1%increase in population,urbanization,energy consumption,per capita industrial GDP,and industrial openness,carbon emissions will increase by 1.04%,0.81%,0.38%,0.27%,and 0.18%accordingly.The coefficient of determination of the GWO-LSTM industrial carbon emission prediction model is 0.9968,which is higher than that of the original LSTM prediction model and SVR prediction model.