文献类型：期刊文献

中文题名：上市公司财务异常与舞弊疑点检测研究

英文题名：Financial Anomalies and Fraud Doubts Detection of Listed Companies

作者：徐静[1];李俊林[1];唐少清[1]

第一作者：徐静

机构：[1]北京联合大学管理学院,北京100101

第一机构：北京联合大学管理学院

年份：2021

期号：S01

起止页码：421-428

中文期刊名：中国软科学

外文期刊名：China Soft Science

收录：CSTPCD;;北大核心:【北大核心2020】；CSCD:【CSCD2021_2022】；CSSCI:【CSSCI2021_2022】；

基金：北京市社会科学基金规划项目—大数据审计模式下财务报表审计线索发现研究(21GLB015)。

语种：中文

中文关键词：财务异常;舞弊疑点;特征选择;异常检测;数据挖掘

外文关键词：financial anomalies;fraud doubts;feature selection;anomaly detection;data mining

摘要：对海量数据价值的挖掘趋势迎来了一个前所未有的大数据时代,将数据挖掘技术应用于财务舞弊疑点发现研究成为一个重要的新兴课题。选取2016—2020年我国深沪两市A股医药制造业上市公司为研究对象,以表征财务报表舞弊的关键指标作为检测变量,分别基于跨年数据和年度数据进行异常检测,从中挖掘出存在异常值的离群点上市公司,进而结合违规处罚信息对其进行疑点验证。研究结果表明:跨年数据样本量更大,模型学习更充分,异常检测的效果优于小样本的年度数据;分析导致异常的原因发现,离群点上市公司表现出资产负债率较高、总资产净利率偏低的特点,为识别可能的财务报表舞弊疑点提供了线索。本研究有助于科学界定重点监管的范围,精准锁定舞弊疑点对象,从而为监管部门甄别财务报表舞弊提供决策支持,对于构建现代化监管执法新模式具有重要的意义。
The trend of mining values of massive data ushers in an unprecedented era of big data.The application of data mining technology in financial fraud detection has become an important new topic.This paper selects A-share pharmaceutical manufacturing listed companies in Shenzhen and Shanghai stock exchanges from 2016 to 2020 as the research object,takes the key indicators of financial statement fraud as the detection variables,and conducts anomaly detection based on cross year data and annual data respectively,so as to find outlier listed companies with abnormal data,and then verifies the suspicious points according to punishment information.The results show that:the sample size of cross year data is larger,the model learning is more sufficient,and the effect of anomaly detection is better than that of small sample annual data;By analyzing the causes of the anomalies,it is found that the outlier listed companies show the characteristics of high asset liability ratio and low net interest rate of total assets,which provides clues for identifying possible financial statement fraud.This study helps to scientifically define the scope of key supervision,accurately lock in the fraud doubts,so as to provide decision support for the regulatory authorities to screen fraud,which is of great significance for building a new mode of modern supervision and law enforcement.