T-S型FNN在公司财务危机预警中的应用  

Application of fussy neural network in financial crisis warning

作　　者：孙爱香[1] 

机构地区：[1]山东理工大学商学院,山东淄博255012

出　　处：《山东理工大学学报（自然科学版）》2017年第4期52-56,共5页Journal of Shandong University of Technology:Natural Science Edition

基　　金：国家社会科学基金青年项目(13CTQ023)

摘　　要：建立一套预测精度高的预警系统,能在公司财务危机出现之前敲响警钟是一个亟待解决的问题.以往的基于统计学方法财务危机预警系统要求输入数据符合一定的统计假设,导致预警系统预测精度偏低,文中选择了另一种核心算法——T-S型FNN来构建公司的财务危机预警系统,将模糊系统和人工神经网络结合起来,使得它的学习能力和表达能力得到了很大提高.实验结果表明:与T-S模糊预警系统、单纯的BP人工神经网络预警系统、MDA(多元判别分析)统计学预警系统相比,T-S型FNN财务危机预警系统的预测精度有了很大提高.The company must establish a high prediction accuracy of early warning system,and it must sound the alarm before crisis.The previous financial crisis early-warning system based on statistical method requires the input data meeting certain statistical assumptions,so it′s forecasting accuracy is very low.In this thesis,a core algorithm—T-S fuzzy neural network is chosen to build the company′s financial crisis early warning system.The T-S fuzzy neural network absorbs the advantages of fuzzy system and neural network,organicly combinated the two method,so the early-warning system′s learning ability and expression ability have been greatly improved.The experimental results show that：the prediction accuracy of the T-S fuzzy neural network financial crisis early-warning system is greatly improved compared to the BP ANN financial crisis early-warning system,the MDA financial crisis early-warning system,and the T-S fuzzy financial crisis early-warning system.

关 键 词：财务危机 FNN T-S 预警模型 

分 类 号：F812[经济管理—财政学]