节点文献

机器学习如何赋能管理学研究?——国内外前沿综述和未来展望

How Can Machine Learning Empower Management Research? A Domestic-Foreign Frontier Review and Future Prospects

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 刘景江郑畅然洪永淼

【Author】 Liu Jingjiang;Zheng Changran;Hong Yongmiao;School of Management, Zhejiang University;School of Economics and Management, University of Chinese Academy of Science;MOE Social Science Laboratory of Digital Economic Forecasts and Policy Simulation, University of Chinese Academy of Science;Academy of Mathematics and System Science, Chinese Academy of Sciences;

【通讯作者】 洪永淼;

【机构】 浙江大学管理学院中国科学院大学经济与管理学院中国科学院大学数字经济监测预测预警与政策仿真教育部哲学社会科学实验室中国科学院数学与系统科学研究院

【摘要】 机器学习正在深刻改变管理学的研究范式与方法。如何运用机器学习更好地赋能管理学研究已经成为学术界关注的前沿热点议题。然而,机器学习在中国管理学研究中的应用仍处于初级阶段。本文基于1999~2021年发表在工商管理和会计财务两大研究领域的国内外顶级期刊的学术文献,识别了学术界借助机器学习开展管理学实证研究的4种核心途径:变量测量、事件预测(包括事件分类)、因果推断和理论构建;梳理了每个途径的代表性文献的研究主题、研究问题、数据集、机器学习算法和研究结论;提出了使用机器学习赋能管理学研究的主要策略,并讨论了中国学者运用机器学习开展中国特色管理理论研究的未来机会。本文显示:将机器学习与传统计量经济学相结合有助于做出更加精准的因果推断;机器学习能够在模式发现这一理论构建的关键步骤中发挥重要作用;将机器学习与多案例分析相结合有助于富有成效地开展理论构建。本文为如何采用机器学习提升管理学研究质量、推进管理学研究范式变革和构建中国特色管理理论提供了方法论指引和方向性启示。

【Abstract】 Machine learning is profoundly changing the research paradigm and method of management. How to use machine learning to empower management research better has become a cutting-edge hot topic of academic attention. However, the application of machine learning in Chinese management research is still in its early stages. Based on academic literature published in top domestic and international journals in the fields of business management, accounting, and finance from 1999 to 2021, this paper identifies four core approaches for the academic community to conduct empirical research in management through machine learning:variable measurement, event prediction(including event classification), causal inference, and theory building. It then sorts out the research topics, questions, data sets, machine learning algorithms, and research conclusions of the representative literature for each approach. It also puts forward the main strategies for empowering management research using machine learning. Finally, it discusses future opportunities for Chinese scholars to use machine learning to study management theory with Chinese characteristics. This paper demonstrates that: Combining machine learning with traditional econometrics can help make more accurate causal inferences; Machine learning can play an important role in pattern discovery,a critical step in theory building; Combining machine learning with multi-case analysis enables us to carry out theory building effectively.Overall, it provides methodological guidance and directional inspiration for how to use machine learning to enhance the quality of management research, promote the paradigm change of management research, and build management theory with Chinese characteristics.

【基金】 国家自然科学基金基础科学中心项目“计量建模与经济政策研究”(基金号:71988101)、国家自然科学基金项目“绩效反馈、制度逻辑与科技企业创业成长研究”(基金号:71872164)和国家自然科学基金项目“患者偏差行为的形成机制与医患关系管理策略研究”(基金号:72074189)的资助
  • 【文献出处】 管理世界 ,Journal of Management World , 编辑部邮箱 ,2023年09期
  • 【分类号】C93;F233;TP181
  • 【下载频次】2887
节点文献中: 

本文链接的文献网络图示:

本文的引文网络