统计与决策2024，Vol.40Issue(1)：52-56,5.DOI:10.13546/j.cnki.tjyjc.2024.01.009
计量经济学中的机器学习方法:回顾与展望
Machine Learning Methods in Econometrics:Retrospect and Prospect
摘要
Abstract
This paper starts with the three major functions of traditional econometrics—economic forecasting,verification theory and policy evaluation,and then sorts out the application of machine learning in econometrics in the process of the integra-tion of machine learning and econometrics.The results go as the following:Machine learning can more efficiently explore the rela-tionship between variables to make more accurate predictions.In the context of big data,machine learning enriches the diversity of economic data.By analyzing and processing a large amount of economic data,it can reveal the potential patterns,correlations and trends in the data,and then test some more complex and challenging research hypotheses.Machine learning improves the accuracy and stability of causal effect estimates from the perspectives of identifying assumptions,functional form fitting,and"counterfactu-al"predictions,so as to quantitatively assess policy effects more accurately and to serve the major strategic needs of the country.关键词
机器学习/计量经济学/大数据/因果推断Key words
machine learning/econometrics/big data/causal inference分类
社会科学引用本文复制引用
石荣,张特,杨国涛..计量经济学中的机器学习方法:回顾与展望[J].统计与决策,2024,40(1):52-56,5.基金项目
国家自然科学基金资助项目(72263027) （72263027）
宁夏哲学社会科学规划重点项目(22NXAYJ02) （22NXAYJ02）
宁夏自然科学基金一般项目(2023AAC03110 （）
2022AAC03027) （）
