基于机器学习的上市公司财务危机预测模型研究  

Research on Financial Crisis Prediction Model of Listed Companies Based on Machine Learning

作　　者：王帅 李卫东[1,2] 张学海 王社伟 李启瑞[1,2] WANG Shuai;LI Weidong;ZHANG Xuehai;WANG Shewei;LI Qirui(College of Information Science and Engineering,Henan University of Technology,Zhengzhou Henan 450001,China;Key Laboratory of Food Information Processing and Control of Ministry of Education,Henan University of Technology,Zhengzhou Henan 450001,China)

机构地区：[1]河南工业大学信息科学与工程学院,河南郑州450001 [2]河南工业大学粮食信息处理与控制教育部重点实验室,河南郑州450001

出　　处：《信息与电脑》2021年第13期49-52,共4页Information & Computer

基　　金：河南省科技开放合作项目(项目编号:182106000025);粮食信息处理与控制教育部重点实验室开放基金(项目编号:KFJJ-2017-101);河南省属高校基本科研业务费专项资金(项目编号:2016XTCX04)。

摘　　要：本文基于公司财务危机预测的6个常用机器学习模型,即逻辑回归模型、K最邻近模型、支持向量机(Support Vector Machine,SVM)模型、朴素贝叶斯模型、决策树模型以及随机森林模型,从东方财富数据中心爬取近两年上市公司财务数据以及股权信息。通过数据预处理,PCA降维,模型参数优化进行多模型性能对比分析,结果显示支持向量机模型具有较高的预测精度,其准确率达到了88%,接受者操作特性曲线(Receiver Operating Characteristic Curve,ROC)曲线面积达到0.89,支持向量机模型对财务危机预测有较好识别能力,这说明支持向量机模型在财务危机预测中具有不错的应用前景和实用价值。Based on six commonly used machine learning models for corporate financial crisis prediction,namely logistic regression model,k-nearest neighbor model,support vector machine model,naive Bayesian model,decision tree model and random forest model,this paper crawls the financial data and equity information of Listed Companies in recent two years from Oriental Wealth data center.Through data preprocessing,PCA dimensionality reduction and model parameter optimization,the performance of multiple models is compared and analyzed.The results show that the support vector machine model has high prediction accuracy,the accuracy rate reaches 88%,the receiver operating characteristic curve(ROC)curve area reaches 0.89,and the support vector machine model has good recognition ability for financial crisis prediction,This shows that support vector machine model has good application prospect and practical value in financial crisis prediction.

关 键 词：财务危机预测 机器学习 预测性能 

分 类 号：F275[经济管理—企业管理]