基于大数据驱动的地方经济短期预测研究

Research on short-term prediction of local economy based on big data

为了进一步提高地方经济短期预测的性能,结合大数据支持,采用深度置信网络深度学习算法对地方经济进行短期预测研究。首先提取地方经济预测分析强相关指标,采用爬虫程序快速有效地通过互联网获取指标数据,借助Python平台将指标自动进行表格输出,将输出数据进行归一化和向量化处理,以地方GDP为预测对象,通过DBN深度学习算法完成地方经济短期预测。在DBN训练过程中,进行权重参数的正反向交叉调节,获得最优参数后确定DBN结构。以重庆市为例进行实例仿真实验表明,基于大数据驱动和DBN深度学习的地方经济短期预测准确度高且MSE小。

In order to further improve the performance of short-term prediction of local economy,combined with big data support,the deep learning algorithm of deep trust network is used to study the short-term prediction of local economy.Firstly,the strong correlation indicators of local economic forecasting analysis are extracted,and the crawler program is used to quickly and effectively obtain the indicator data through the Internet.By means of Python platform,the indicators are automatically output in tables,and the output data is normalized and vectorized.Taking local GDP as the prediction object,the short-term prediction of local economy is completed by DBN deep learning algorithm.In the training process of DBN,the weight parameters are cross-adjusted forward and backward,and the structure of DBN is determined after obtaining the optimal parameters.Taking Chongqing as an example,the simulation experiment shows that the local economy shortterm prediction based on big data driving and DBN deep learning has high accuracy and small MSE.