摘要
本文以ST状况为目标变量的财务困境判别模型表明,对识别企业是否陷入财务困境的最有影响的因素是资产收益率、股东权益比率和净资产收益率增长率指标,并以神经网络技术建立财务困境判别模型,模型评价显示,模型的正确判别率高达90%以上;最后,得出结论和政策建议。当然随着数据环境的改变,必须不断地对其修正和完善,才能适应新的情况。
The paper makes the type of ST situation discriminative models by means of data mining. We find out the most influential factor to distinguish financial difficulty is such two indicators as rate of return on total asset and stockholders＇ equity ratio. The models of financial difficulty distinction set up by BP neural network is good as high as above 90 percentages ;The article puts forward the conclusions and policy suggestions finally. Of course, the models we created are dependent on datum. Along with the variation of data environment, it is necessary to correct and perfect them to adapt to new situation.
出处
《兰州商学院学报》
CSSCI
2010年第2期100-103,共4页
Journal of Lanzhou Commercial College
基金
国家自然科学基金资助项目"基于BP神经网络的上市公司财务困境预警模型实证研究"(70671025)
2009年安徽省高等学校自然科学研究项目"基于BP神经网络的上市公司财务困境预警模型实证研究"(KJ2009B130Z)阶段性研究成果
关键词
财务困境
神经网络
模型评估
financial difficulty
neural network
model assessment
