基于人工神经网络进行财务危机预警的改进方法  

IMPROVED METHOD FOR PREDICTING FINANCIAL CRISIS BASED ON ARTIFICIAL NEURAL NETWORK

作　　者：张承翔[1] 丁鹏[1] 盛焕烨[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《计算机应用与软件》2008年第8期147-148,157,共3页Computer Applications and Software

摘　　要：利用Ruck的研究成果,对人工神经网络进行财务危机预警的方法进行了改进,输出公司未来的财务状况是正常还是陷入危机的概率。利用我国上市公司1998年到2003年的财务数据,实验证明改进方法能够有效预测公司财务状况,提前2年判断公司是否会陷入财务危机的正确率达到82%。Using the research outcome of Ruck, we improved the scheme of ANN-based financial crisis prediction, thus the output becomes the probability of whether the future financial situation in a company would be in normal or with crisis. By using the financial data of Chinag listed companies between 1998 and 2003 ,the experiment proved that the improved scheme can effectively predict the financial situation of a company, and the accuracy rate of judging whether or not the company may falling in financial crisis 2 years ahead achieves 82%.

关 键 词：人工神经网络 财务危机预警 概率 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] F275[自动化与计算机技术—控制科学与工程]