基于LSTM神经网络的内蒙古自治区GDP预测
武阳 , 罗季康 , 赵贞 , 谢晓波 , 庞晶
内蒙古工业大学学报(自然科学版) ›› 2024, Vol. 43 ›› Issue (04) : 296 -301.
基于LSTM神经网络的内蒙古自治区GDP预测
GDP prediction of Inner Mongolia Autonomous Region based on LSTM neural network
通过使用国家统计局公开发布的内蒙古自治区1992—2022年的年度GDP数据,基于长短时记忆神经网络(LSTM)分别构建了两步预测模型和三步预测模型进行对比,并在网络结构中添加了Dropout模块,避免出现过拟合的情况,同时提高模型的预测能力。根据预测值和真实值的结果,使用平均绝对误差(MAE)、均方误差(MSE)、均方根误差(RMSE)指标来评价两个模型的优劣。经实验结果表明:LSTM两步预测模型在测试集样本中R2值可达到0.93,证明该模型预测结果拟合更好,有很强的泛化能力,可用于内蒙古自治区GDP的短期预测;应用LSTM两步预测模型预测内蒙古自治区2023—2024年的GDP值分别为24 805.60亿元和25 131.69亿元,能够看出该地区未来GDP增长良好,可为政府部门定制宏观经济计划提供参考。
The paper uses the annual GDP data of Inner Mongolia Autonomous Region from 1992 to 2022 publicly released by the National Bureau of Statistics. Based on the Long Short Term Memory Neural Network (LSTM), a two-step prediction model and a three-step prediction model are constructed for comparison, and a Dropout module is added to improve the model's generalization ability. Average absolute error (MAE), mean square error (MSE) and root mean square error (RMSE) are used to evaluate the advantages and disadvantages of the two models. The experimental results show that the LSTM two-step prediction model can achieve an R2 value of 0.93 in the test set samples, proving that the model has better fitting and strong generalization ability, and can be used for short-term prediction of GDP for Inner Mongolia Autonomous Region. The LSTM two-step prediction model is used to predict the GDP values of the Inner Mongolia Autonomous Region from 2023 to 2024 to be 2 480.56 billion yuan and 2 513.169 billion yuan respectively. It can be seen that the future GDP growth of the region is good, which can provide a scientific reference for the government to customize the macro economic planning.
| [1] |
马静雯, 李树青, 夏梦瑶. 机器学习在GDP预测中的应用研究述评[J]. 科技情报研究, 2022, 4(3): 73-94. |
| [2] |
李瑞阁, 黄佳艳. 基于ARIMA乘积模型的国民经济GDP预测研究[J]. 南阳理工学院学报, 2018, 10(2): 7-12. |
| [3] |
蒋延军, 李运德, 李树德, |
| [4] |
|
| [5] |
|
| [6] |
梁娜, 张吉刚. 基于改进BP的GDP预测研究[J]. 计算机与数字工程, 2012, 40(1): 49-50, 53. |
| [7] |
朱青, 周石鹏. 基于LSTM-XGBoost二维组合模型的GDP增速预测[J]. 软件导刊, 2021, 20(4): 57-62. |
| [8] |
梁龙跃, 陈玉霞. 基于深度学习神经网络的季度GDP预测[J]. 统计与决策, 2023, 39(2): 24-29. |
| [9] |
|
| [10] |
周志华. 机器学习[M]. 北京: 清华大学出版社, 2016. |
| [11] |
肖争艳, 刘玲君, 赵廷蓉, |
| [12] |
李航. 统计学习方法[M]. 北京: 清华大学出版社, 2012. |
内蒙古自治区直属高校基本科研业务费项目(JY20220003)
/
| 〈 |
|
〉 |