基于循环神经网络的GDP预测研究与分析
Research and analysis of GDP prediction based on cyclic neural network
白斌丽 1吴年祥1
作者信息
- 1. 安徽国防科技职业学院,安徽六安 237011
- 折叠
摘要
GDP(Gross Domestic Product)和人均GDP是一个国家经济实力的标志性指标,反映一个国家经济发展状况.通过世界银行提供各国1976年以来的GDP和人均GDP数据对LSTM(Long Short-Term Memo-ry)网络进行了训练,用训练好的LSTM网络对6个国家的人均GDP进行了预测.通过对预测值和实际值的比较,结果显示LSTM网络对人均GDP的预测效果明显优于传统的统计学方法.
Abstract
GDP(Gross Domestic Product)and GDP per capita are two symbolic indicators of a country's economic strength,reflecting the economic development of a country.The LSTM(Long Short-Term Memory)network are trained through the GDP and per capita GDP data provided by the World Bank since 1976,the result of which is used to predict the per capita GDP of six countries.By compa-ring the predicted value with the actual value,this paper reaches the conclusion that the prediction effect of LSTM network on GDP per capita is significantly better than that by using the traditional statistical method.
关键词
人均GDP/深度学习/循环神经网络/长短期记忆网络Key words
per capita GDP/deep learning/RNN/LSTM引用本文复制引用
基金项目
高等学校拔尖人才项目(gxbjZD2021038)
安徽省教育厅自然科学研究重点项目(KJ2021A1494)
出版年
2024
NETL
NSTL
万方数据