基于灰色关联度的GM(1,1)-BP神经网络组合模型研究

Study on GM( 1,1)-BP Combination Model Based on Grey Correlation

改革开放以来,我国居民消费水平保持较快发展态势,准确预测未来居民消费水平能更好地掌握国民经济发展趋势,为政府相关部门制定战略规划和产业政策提供参考。为此,基于灰色关联度建立了GM(1,1)-BP神经网络组合模型,对2000—2016年居民消费水平进行模拟,并对比分析GM(1,1)模型、BP神经网络模型、灰色组合模型模型的预测误差。仿真结果表明,基于灰色关联度的GM(1,1)-BP神经网络组合模型能进一步提高预测准确性。

Since reform and opening up,the consumption of resident in China has maintained a relatively rapid development,and accurately predicting the future consumption of resident can grasp the trend of national economic development,and provide a reference for government departments to formulate strategic planning and industrial policies.Based on the grey correlation,this paper establishes the GM(1,1)-BP combination model,simulates the resident consumption from 2000 to 2016,and compares the prediction error of GM(1,1)model,BP neural network model and combination model;The results show that the GM(1,1)-BP combination model based on grey correlation can further improve the prediction accuracy.