摘要
目前针对财务危机的研究很多,多数采用了静态财务指标结合机器学习进行预测。针对财务危机是一个动态演变的过程这个问题,考虑到财务指标在时间上的动态变化在一定程度上反映公司的财务状况,提出了基于静态财务指标的动态财务指标。同时针对财务欺诈对于财务危机判断的影响的问题,加入了反欺诈指标,并与深度神经网络框架结合,建立财务危机预测模型。通过实验对提出的模型进行准确率测试。实验结果表明,增加的动态财务指标能有效提高预测模型的准确率,结合深度神经网络的模型获得了最好的预测准确率,达到89.97%。增加的反欺诈指标能有效检测财务欺诈的样本,增强预警模型的现实意义。
At present, there are many researches on financial crisis, and most of them use static financial indicators combined with machine learning for prediction. In view of the fact that financial crisis is a dynamic evolution process, considering the dynamic changes of financial indicators in time to reflect the financial status of the company to a certain extent, a dynamic financial indicator based on static financial indicators was proposed. At the same time, anti-fraud indicators were added to address the impact of financial fraud on financial crisis judgments, and deep neural network framework was used to establish a financial crisis prediction model. The accuracy of the proposed model was tested by experiments. The experimental results show that the increased dynamic financial indicators can effectively improve the accuracy of the prediction model. The model with deep neural network has obtained the best prediction accuracy, reaching 89.97%. The increased anti-fraud indicators can effectively detect the sample of financial fraud and enhance the practical significance of the early warning model.
作者
李辰杰
LI Chenjie(College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou Zhejiang 310023,China)
出处
《计算机应用》
CSCD
北大核心
2019年第S02期275-279,共5页
journal of Computer Applications
关键词
动态指标
财务危机
深度神经网络
反欺诈
dynamic indicator
financial crisis
Deep Neural Network(DNN)
anti-fraud
