基于灰色关联度分析和GRU神经网络的辽宁省GDP预测

GDP Prediction of Liaoning Province Based on Grey Correlation Analysis and GRU Neural Network

摘要: 利用灰色关联分析方法分析了影响辽宁省GDP的宏观经济指标，选取相关性较高的指标输入到门控循环单元神经网络中，建立基于灰色关联分析的门控循环单元模型。首先利用灰色关联分析计算10个宏观经济指标与生产总值之间的相关度，然后根据计算结果，选取出5个关联度最大的指标建立灰色关联-GRU、灰色关联-LSTM和单一的GRU神经网络模型以及SVM模型对辽宁省生产总值进行预测，将模型的预测结果进行比较，结果表明基于灰色关联分析和GRU模型的预测精度较高。

Abstract: The macroeconomic indicators that affect GDP of Liaoning province are analyzed by using grey correlation analysis method, the indicators with high correlation with GDP are selected and input into the gated recurrent neural network, and the gated recurrent neural network model based on grey correlation analysis is established. Firstly, the correlation between 10 macroeconomic indicators and GDP is calculated by grey correlation analysis, then five indexes with strong correlation from several macroeconomic indicators are selected to establish GRU neural network based on grey correlation analysis, LSTM neural network based on grey correlation analysis, GRU neural network and SVM to predict the GDP of Liaoning Province. Comparing the prediction results of these models, the results show that the GRU neural network based on grey correlation analysis has high prediction accuracy.