虚假财务报表的FCM算法识别  

Application of FCM algorithm in fraudulent financial information

作　　者：吴杰[1] 耿新青[2] 

机构地区：[1]鞍山市康宁医院财会科,辽宁鞍山114001 [2]鞍山师范学院数学与信息科学学院,辽宁鞍山114001

出　　处：《辽宁科技大学学报》2016年第4期307-310,共4页Journal of University of Science and Technology Liaoning

摘　　要：财会人员借助财务惯例和财务准则等财务技术手段篡改财务报表,与常规财务数据对照会发现,所篡改的财务数据在财务数据指标方面往往缺少相互关联性,甚至是自相矛盾。同时在报表模式上与常规财会数据也呈现内在差别,但是这些差别人工检验较难。本文采用模糊C均值聚类方法将这些差异放大,从而识别出虚假财务报表。该方法具有识别准确率高和快速的特点,具有理论方法创新意义。The pattern identification algorithm applied in fraudulent financial information is of importance in theory innovation. The accountants fabricate the financial data with the traditional method, which follows accounting convention and accounting standard, also including the profession of certified public accountant.Compared with natural data, the fraudulent financial data show the defects of data structure. The financial information is provided with inherent difference, which is not notable in the opinion of the accountants. The difference is magnified with statistics or data mining technique. The FCM algorithm is used to clustering analysis in fraudulent financial information. The method has the characteristics of high accuracy, high efficiency, and the significance of theory innovation.

关 键 词：虚假财务报表 会计舞弊 数据挖掘 

分 类 号：F239[经济管理—会计学]