文献类型：期刊文献

中文题名：基于数据挖掘的重大错报风险识别和评估研究

英文题名：Identification and Assessment of Risk of Material Misstatement Based on Data Mining

作者：徐静[1];李俊林[1]

第一作者：徐静

机构：[1]北京联合大学管理学院,北京100101

第一机构：北京联合大学管理学院

年份：2022

卷号：43

期号：6

起止页码：79-85

中文期刊名：财经理论与实践

外文期刊名：The Theory and Practice of Finance and Economics

收录：CSTPCD;;国家哲学社会科学学术期刊数据库;北大核心:【北大核心2020】；CSSCI:【CSSCI2021_2022】；

基金：北京市社会科学基金规划项目(21GLB015)。

语种：中文

中文关键词：数据挖掘;重大错报风险;CHAID算法;分类预测;决策规则;

外文关键词：data mining;risk of material misstatement;CHAID algorithm;classification prediction;decision rule;

摘要：基于现代风险导向审计理论,运用数据挖掘方法,选取2001-2020年因财务报表重大错报被出具保留和否定意见的上市公司为样本,构建基于CHAID算法的重大错报分类预测模型,从中识别重大错报的基本特征,挖掘重大错报与否的决策规则。结果表明:流动性差、盈利能力不足、长期偿债能力弱的上市公司,在较高的置信水平上存在高重大错报风险;模型总体正确率为97.17%,分类效果较为理想。鉴于此,能够为识别、评估和预测财务报表重大错报风险提供线索。
Based on the modern risk-oriented auditing theory and data mining method,taking the listed companies with reservations and negative opinions due to material misstatement from 2001 to 2020 as the sample,a classification and prediction model of material misstatement based on CHAID algorithm is built.From which it’s able to identify the basic characteristics of material misstatement,and mine the decision-making rules of whether or not to make material misstatement.The results show that listed companies with poor liquidity,insufficient profitability and weak long-term solvency show high risk of material misstatement(RMM)at a high confidence level.And as well,the overall accuracy of the model is 97.17%,which indicates that the classification effect is ideal.In this regard,clues for identifying,assessing and predicting the RMM of financial statements are provided.