摘要
建立精准的舞弊识别模型是个难题,因此相关研究从未中断,已有的舞弊识别的二分类模型,其构建基础是舞弊和非舞弊样本公司都是确定的,但实际情况是受到惩罚的舞弊公司是确定的而非舞弊公司难以判断,造成分类模型的构建基础不复存在、识别准确率低,此时,构建会计舞弊指数便是很好的替代。舞弊指数的构建基础是假设每个公司都有舞弊的可能,而无须事先知道舞弊与否的更多信息。指数具有连续、灵敏、精细化刻画舞弊程度的优点。除了探讨了二分类Logistic回归会计舞弊识别模型的原理、构建基础和优缺点外,还探讨了会计舞弊指数的构建方法、构建基础和优缺点,为开发会计舞弊识别模型和会计舞弊指数提供了极具价值的方法借鉴。
The related research for establishment of accurate fraud identification model has never been interrupted because it is a hard problem.Establishment foundation for the fraud identification model on two classification logistic is that the fraud and non fraud sample firms selected are determined,but as a matter of fact,fraud companies that are punished are determined,but non fraud companies are difficult to judge,which leads to false model construction foundation and low recognition rate,so the construction of accounting fraud index is a good alternative,which is based on the assumption that every company has the possibility of fraud,more information about whether it is fraud or not is not needed in advance.The index has the advantages of continuous,sensitive and precise description of fraud degree.In addition to regression principle,establishment foundation,advantages and disadvantages of accounting fraud identification model on two classification logistic are referred,construction method,establishment foundation,advantages and disadvantages for accounting fraud index are focused in the paper,providing a reference for the development of a valuable method of accounting fraud identification model and accounting fraud index.
作者
李清
闫世刚
LI Qing;YAN Shi-gang(Business School,Jilin University,Changchun 130012,China)
出处
《东北师大学报（哲学社会科学版）》
CSSCI
北大核心
2018年第4期176-182,共7页
Journal of Northeast Normal University(Philosophy and Social Science Edition)
基金
国家社科基金项目(18BGL067)
吉林省社科基金项目(2018B84)
吉林省教育厅"十三五"社会科学规划项目(JJKH20170882SK)
