A Research on Listed Companies Financial Pre-warning Model Based on BP Neural Network

摘要：

Abstracts:Nowadays,bankruptcy is a regular phenomenon under fierce competitive circumstances.Scholars have done many researches on financial pre-warning models.The studies evolved from single-variable analysis model to multiple-variable model,and finally to build model by using statistics methods.However,these classical models, such as LO GISTIC Model and PROBIT Model,still have disadvantages,for exam ple they have multicollinearity sensitivity,outlier sensitivity and missing value sensitivity.In order to conquer these disadvantages,this article ap plies BP neural network into the financial prewarning model.It looks into Chinese listed companies information from 2005 to 2007.It selects 55 ST companies as samples,and matches 80 normal companies in the same year,the same industry and of similar size.It uses significant test to filter the index,and finally chooses 13 indexes to be input nodes of the mod el.After training and studying the sample,the models result shows that the effect of financial pre-warning model based on BP neural network works well.

关键词： BP neural network Financial distress Financial pre-warning model Listed companies

作者: Ding Lin

作者单位: School of Accounting,Zhongnan University of Economies and Law

会议类型: 国际会议

会议名称: 2008 Sino-Australia International Conference on Accounting and Finance(2008中澳会计与财务国际学术研讨会)

会议地点: 武汉

会议语种:英文

页码: 560-570

在线出版日期: 2008-10-31（万方平台首次上网日期，不代表论文的发表时间）