摘要
旨在大数据框架下,探讨利用机器学习算法筛选出刻画公司财务舞弊与财务异常行为预警的特征指标及相关应用,即,我们首先筛选可描述公司财务舞弊与财务异常行为的相关关联特征指标,然后以这些指标为基础,进行关联特征选取并针对其预警或解释能力进行甄别测试。具体来讲,通过刻画财务舞弊的特征指标和基于非结构化的公司治理结构的风险特征的提取,建立解读公司治理结构(特别是针对财务舞弊行为等方面)的风险预警体系需要的关联特征指标,并将这些指标应用于真实案例的测试分析。我们的案例分析表明,从财务到公司治理框架层面构建的针对财务舞弊行为和财务异常状态的特征指标刻画可以达到预测财务舞弊和预警的目的。文章的创新之处在于:基于公司舞弊三角原理,尝试采用大数据特征提取方法,能够构建刻画公司财务质量和公司治理结构的财务舞弊风险评估体系需要的关联特征因子指标;同时我们也期待支持该体系的特征指标可以对公司可能要发生的财务舞弊进行及时的预警,从而促使行业健康发展并避免欺诈等不良行为带来的潜在损失。
This paper mainly introduces how to use the machine learning algorithm based on artificial intelligence under the big data framework to establish and interpret the corporate governance structure(especially corporate financial fraud)by depicting the indicators of financial fraud and extracting the risk characteristics based on unstructured corporate governance structure Our case analysis shows that the characterization indicators for the outbreak of financial fraud and financial anomalies constructed from the financial to corporate governance framework are predicted in advance by the ability to achieve the purpose of early warning;especially,based on the fraud triangle principle,the big data feature extraction method is aimed at the corporate finance The establishment of a comprehensive evaluation system for quality and corporate governance structure can establish a timely early warning and prevention against the outbreak of financial fraud and financial abnormalities,so as to achieve the healthy development of the industry and avoid potential losses caused by financial fraud.
作者
袁先智
周云鹏
严诚幸
王一伊
何华
张启珑
Yuan Xianzhi;Zhou Yunpeng;Yan Chengxing;Wang Yiyi;He Hua;Zhang Qilong(School of Economics and Finance,Chongqing University of Technology,Chongqing 400054,China;Business School,Sun Yat‑sen University,Guangzhou 510275,China;Shanghai Hammer Digital Technology Co.,Ltd.,Shanghai 200093,China;Beijing 2022 Winter Olympic and Paralympic Organizing Committee,Beijing 102022,China)
出处
《数智技术研究与应用》
2025年第1期28-38,共11页
SmartTech Innovations
基金
国家自然科学基金资助项目(U1811462、71971031)。
