Danyang Wei, Soohyun Cho, Miklos V. Vasarhelyi, Liam Te-Wierik
{"title":"审计中的异常值检测:将无监督学习融入总账分析的多层次框架中","authors":"Danyang Wei, Soohyun Cho, Miklos V. Vasarhelyi, Liam Te-Wierik","doi":"10.2308/isys-2022-026","DOIUrl":null,"url":null,"abstract":"\n Auditors traditionally use sampling techniques to examine general ledger (GL) data, which suffer from sampling risks. Hence, recent research proposes full-population testing techniques, such as suspicion scoring, which rely on auditors’ judgment to recognize possible risk factors and develop corresponding risk filters to identify abnormal transactions. Thus, when auditors miss potential problems, the related transactions are not likely to be identified. This paper uses unsupervised outlier detection methods, which require no prior knowledge about outliers in a dataset, to identify outliers in GL data and tests whether auditors can gain new insights from those identified outliers. A framework called the Multilevel Outlier Detection Framework (MODF) is proposed to identify outliers at the transaction level, account level, and combination-by-variable level. Experiments with one real and one synthetic GL dataset demonstrate that the MODF can help auditors to gain new insights about GL data.\n Data Availability: The real dataset used in the experiment is not publicly available due to privacy policies.\n JEL Classifications: M410, M42.","PeriodicalId":46998,"journal":{"name":"Journal of Information Systems","volume":null,"pages":null},"PeriodicalIF":2.0000,"publicationDate":"2024-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Outlier Detection in Auditing: Integrating Unsupervised Learning within a Multilevel Framework for General Ledger Analysis\",\"authors\":\"Danyang Wei, Soohyun Cho, Miklos V. Vasarhelyi, Liam Te-Wierik\",\"doi\":\"10.2308/isys-2022-026\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\n Auditors traditionally use sampling techniques to examine general ledger (GL) data, which suffer from sampling risks. Hence, recent research proposes full-population testing techniques, such as suspicion scoring, which rely on auditors’ judgment to recognize possible risk factors and develop corresponding risk filters to identify abnormal transactions. Thus, when auditors miss potential problems, the related transactions are not likely to be identified. This paper uses unsupervised outlier detection methods, which require no prior knowledge about outliers in a dataset, to identify outliers in GL data and tests whether auditors can gain new insights from those identified outliers. A framework called the Multilevel Outlier Detection Framework (MODF) is proposed to identify outliers at the transaction level, account level, and combination-by-variable level. Experiments with one real and one synthetic GL dataset demonstrate that the MODF can help auditors to gain new insights about GL data.\\n Data Availability: The real dataset used in the experiment is not publicly available due to privacy policies.\\n JEL Classifications: M410, M42.\",\"PeriodicalId\":46998,\"journal\":{\"name\":\"Journal of Information Systems\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.0000,\"publicationDate\":\"2024-05-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Journal of Information Systems\",\"FirstCategoryId\":\"91\",\"ListUrlMain\":\"https://doi.org/10.2308/isys-2022-026\",\"RegionNum\":4,\"RegionCategory\":\"管理学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"BUSINESS, FINANCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Information Systems","FirstCategoryId":"91","ListUrlMain":"https://doi.org/10.2308/isys-2022-026","RegionNum":4,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"BUSINESS, FINANCE","Score":null,"Total":0}
Outlier Detection in Auditing: Integrating Unsupervised Learning within a Multilevel Framework for General Ledger Analysis
Auditors traditionally use sampling techniques to examine general ledger (GL) data, which suffer from sampling risks. Hence, recent research proposes full-population testing techniques, such as suspicion scoring, which rely on auditors’ judgment to recognize possible risk factors and develop corresponding risk filters to identify abnormal transactions. Thus, when auditors miss potential problems, the related transactions are not likely to be identified. This paper uses unsupervised outlier detection methods, which require no prior knowledge about outliers in a dataset, to identify outliers in GL data and tests whether auditors can gain new insights from those identified outliers. A framework called the Multilevel Outlier Detection Framework (MODF) is proposed to identify outliers at the transaction level, account level, and combination-by-variable level. Experiments with one real and one synthetic GL dataset demonstrate that the MODF can help auditors to gain new insights about GL data.
Data Availability: The real dataset used in the experiment is not publicly available due to privacy policies.
JEL Classifications: M410, M42.
期刊介绍:
The Journal of Information Systems (JIS) is the academic journal of the Accounting Information Systems (AIS) Section of the American Accounting Association. Its goal is to support, promote, and advance Accounting Information Systems knowledge. The primary criterion for publication in JIS is contribution to the accounting information systems (AIS), accounting and auditing domains by the application or understanding of information technology theory and practice. AIS research draws upon and is informed by research and practice in management information systems, computer science, accounting, auditing as well as cognate disciplines including philosophy, psychology, and management science. JIS welcomes research that employs a wide variety of research methods including qualitative, field study, case study, behavioral, experimental, archival, analytical and markets-based.