基于改进型神经网络的财务预警实证研究

An Empirical Analysis on Improved-neural Financial Precautionary Networks

传统的神经网络财务预警模型存在缺陷,无法解释变量间的因果关系,使训练时间增加、训练精度下降。改进后的神经网络预警模型运用了统计方法、神经网络、粗糙集和模糊技术,能够得到优于传统神经网络模型的预测结果,弥补了传统神经网络模型解释性差的缺陷,实证研究结果也证实了该模型的有效性。

The paper analyzes the limitations of the traditional neural networks which is extensively used in the financial precaution. The author makes a research on the detailed methods in improving financial precaution networks. The result of the empirical analysis proves the efficiency of the related model as well,