摘要
研究了财务报表舞弊识别问题。以2011-2020年深沪A股上市公司的财务报表为样本数据,引入信息值构建指标筛选模型,提取17个财务变量和4个非财务变量,对样本数据进行清洗和归一化后,运用XGBoost算法对样本数据进行分类。实验结果表明,基于XGBoost算法构建的财务报表舞弊识别模型在所有性能指标上都优于机器学习算法中的逻辑回归、支持向量机和随机森林算法。
In this paper,the identification of financial statement fraud is studied.Taking the financial statements of ShenzhenShanghai A-share listed companies from 2011 to 2020 as sample data,the information value is introduced to construct an index screening model,and 17 financial variables and 4 non-financial variables are extracted.After cleaning and normalizing the sample data,the XGBoost algorithm is used to classify the sample data.The experimental results show that the financial statement fraud identification model based on the XGBoost algorithm is superior to the logistic regression(LR),support vector machine(SVM)and random forest(RF)algorithms in machine learning algorithms in all performance indicators.
作者
吴贞如
Wu Zhenru(School of Information and Engineering,Nanjing Audit University,Nanjing,Jiangsu 211815,China)
出处
《计算机时代》
2022年第8期29-33,共5页
Computer Era