机器学习算法在金融控股公司信息系统审计中的应用探究  

Exploration of application of machine learning algorithm in information system audit of financial holding companies

作　　者：车纲 王振宇 Che Gang;Wang Zhenyu(China Resources Group,Shenzhen,Guangdong,518052;China Resources Pharmaceutical Business Group Co.,Ltd.,Beijing,100009)

机构地区：[1]华润集团,广东深圳518052 [2]华润医药商业集团有限公司,北京100009

出　　处：《市场周刊》2024年第24期88-91,共4页Market Weekly

摘　　要：随着金融行业的数字化转型,信息系统的安全性和效率对金融控股公司至关重要。文章探讨了无监督学习算法,特别是隔离森林算法,在金融信息系统审计中的应用。通过分析具体的安全事件,文章展示了无监督学习算法如何帮助审计师识别异常行为,提前发现潜在的安全漏洞和非法活动;研究提出了将隔离森林算法集成到现有审计流程中的方案,并讨论了集成方法的优势,包括提高审计效率、增强审计准确性和加强数据保护。此外,文章还进行了理论模拟的案例分析,深入探讨了算法在信用卡交易数据中的应用,并提出了未来研究方向;讨论了引入AI算法的挑战,包括数据隐私、技术整合、算法透明度和问责性,并提出了相应的建议。With the digital transformation of the financial industry,the security and efficiency of information systems are crucial for financial holding companies.This paper explores the application of unsupervised learning algorithms,especially the Isolation Forest algorithm,in the auditing of financial information systems.By analyzing specific security incidents,the paper demonstrates how unsupervised learning algorithms can assist auditors in identifying abnormal behaviors and detecting potential security vulnerabilities and illegal activities in advance.The study proposes a framework for integrating the Isolation Forest algorithm into existing audit processes and discusses the advantages of the integrated approach,including improving audit efficiency,enhancing audit accuracy and strengthening data protection.Furthermore,the paper conducts a theoretical simulation case analysis,delving into the application of the algorithm in credit card transaction data,and suggests directions for future research.Finally,the paper discusses the challenges of introducing AI algorithms,including data privacy,technology integration,algorithm transparency,and accountability,and provides corresponding recommendations.

关 键 词：信息系统审计 金融控股公司 无监督学习算法 隔离森林算法 

分 类 号：F239.2[经济管理—会计学]