摘要
Abstract
Traditional linear financial risk warning technology for enterprises can no longer meet the needs of enterprise development.In order to solve the problem of insufficient intelligent financial risk warning,it is necessary to analyze enterprise financial risks and construct a risk indicator system.At the same time,in order to explore the complex relationships between data,Convolutional Neural Networks(CNN)are used to identify risk features,and Long Short Term Memory(LSTM)networks are used to handle the dependency relationship between information reading and information storage.Simultaneously introducing sliding window algorithm to optimize the warning model.In the prediction of ROC results in financial risk prediction,the proposed model has the best warning training effect among financial warning data divided by year,with an AUC value of 0.896.In the risk detection of development ability and repayment ability,the proposed model has better risk detection performance than other models,with accuracy rates of 96.65% and 95.75%,respectively.From this,it can be seen that the proposed risk model has excellent application effects,providing technical references for the application and risk supervision of enterprise intelligent finance.关键词
深度预警/智能财务/风险指标体系/卷积神经网络/长短时记忆网络Key words
Deep warning/Intelligent finance/Risk indicator system/Convolutional network/Long and Short Term Networks分类
管理科学
