我国上市公司财务报告舞弊识别模型对比研究

Comparative study on Financial Reporting Fraud Identification Model of Listing Corporation in China

在我国沪深A股上市公司中,选取年报舞弊的公司45家、健康公司95家,建立了4变量BP神经网络和Logistic回归财务报告舞弊识别模型,结果表明,BP神经网络模型具有更高的识别准确率,总准确率达到了87.10%。

Based on Shanghai and Shenzhen A-share listed companies as research samples in this paper,the study selects 45 fraudulent listed companies and 95 normal listed companies,then establishes BP neural network and logistic regression model of detecting fraudulent financial reports with choosing 4 indices.The results show that BP neural network model with a higher detecting fraud accuracy,and the total accuracy rate of 87.10%.