基于Shapley值的中国能源消费组合预测模型研究

Research on the portfolio prediction methods of China energy consumption based on Shapley value

通过指数平滑模型、系统动力学模型、GM(1,1)模型分别对中国能源消费总量进行预测,进一步运用合作博弈中的Shapley值方法,通过分配总误差来确定组合预测模型中各预测模型的权重,以此构建组合预测模型,并对中国能源消费总量进行预测。预测结果表明,该组合预测模型的预测精度高于其他各分项预测模型,对能源消费的预测是可行、有效的。

Using exponential smoothing model,system dynamics model and gray prediction model,the energy consumption of China was forecasted,the Shapley value method of cooperative game was applied to determine the weight of each forecast model in the combined forecast model by distributing the total combined error,and then the energy consumption of China was forecasted through the combined model.The result showed that the prediction accuracy of the combined model was higher than that of the model selected,and the prediction of energy consumption was feasible and effective.