摘要
笔者针对提高公司财务风险判别准确度问题,提出运用人工神经网络中具有明显自适应性、自组织性的自适应共振理论算法和自组织特征映射算法分别构建公司财务风险判别模型进行仿真研究。结果表明:自适应共振理论算法判别正确率为87%,而自组织特征映射算法判别正确率则达到了89%,较反向传播神经网络等方法判别准确度更高。
This paper aims at finding a method to improve the accuracy degree of company＇s financial risk identification. The author puts forward the adaptive resonance theory algorithm and the self-organizing feature map algorithm,which has obvious features of adaptability and self-organizing,respectively to build the company＇s financial risk discriminant model for the simulation research. The results show that the recognition accuracy of Adaptive Resonance Theory neural network reaches 87%,and Self- organizing Feature Map network algorithm reaches 89%,both of which have better recognition effects than other Artificial Neural Networks such as BP network algorithm,ect.
出处
《经济经纬》
CSSCI
北大核心
2017年第2期122-127,共6页
Economic Survey
基金
国家自然科学基金项目(71561002)
国家民族委员会科研项目(14BFZ020)
国家民族委员会经济管理重点开放实验室项目
宁夏中小企业研究所科研项目(ZXS2015001)
关键词
财务风险
风险判别
神经网络
自适应共振理论
自组织特征映射
Financial Risk
Risk Identification
Neural Network
Adaptive Resonance Theory
Self-organizing Feature Map
