基于数据分类的企业财务数据异常判定方法  

Judgment Method of Enterprise Financial Data Abnormal Based on Data Classification

作　　者：李苗[1] LI Miao(Shanxi Vocational College of Finance and Economics,Xianyang 712000 China)

机构地区：[1]陕西财经职业技术学院,陕西咸阳712000

出　　处：《自动化技术与应用》2023年第10期91-94,共4页Techniques of Automation and Applications

摘　　要：采用目前方法对企业财务数据进行异常判定时,存在AUC值小、相对误差大等问题,提出基于数据分类的企业财务数据异常判定方法。通过模糊聚类方法对原始企业财务数据集进行预处理,依据支持向量机原理构建异常数据判定模型,结合企业财务实际情况通过残差计算完成数据异常的判定。实验结果表明:所提方法能够有效地增大AUC值、减小相对误差、缩短判定时间并降低漏判率。When using the current method to judge the abnormality of enterprise financial data,there are some problems,such as small AUC value and large relative error.A method to judge the abnormality of enterprise financial data based on data classification is proposed.The original enterprise financial data set is preprocessed by fuzzy clustering method,and the abnormal data judgment model is constructed according to the principle of support vector machine.Combined with the actual situation of enterprise finance,the abnormal data judgment is completed by residual calculation.The experimental results show that the proposed method can effectively increase the AUC value,reduce the relative error,shorten the judgment time and reduce the missed judgment rate.

关 键 词：数据分类 企业财务数据 异常判定 聚类处理 支持向量机 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]