Empirical Studies Based On Discriminant Analysis And Neural Network Technology, Listed Companies In Financial Distress Early Warning

Financial distress is a world problem. Since 1960s,studies on corporate failure prediction have prevailed both in the U.S.A and in European countries. But few researches on this subject have been conducted in our country until very recently, most of papers are focus on the quantitative analysis but qualitative analysis. This paper make use of the financial statements of listed companies in China Stock Exchange and build a successful model for corporate failure discrimination by MDA and ANN. The results show that Ann's prediction ability is superior than MDA, both models can predict corporate financial distress very well. This research concludes that under current economic and accounting environment in China, financial statements can sufficiently reflect the characteristics of financial distress corporations. Our study includes four chapters as follows: In chapter 1, we discuss the concept of financial distress and its significance, and also review the important literature on financial distress from domestic and abroad papers. We then discuss the main progress in this study. In chapter 2, we introduce how we design our study. The research methodology in this study is profile analysis, MDA and ANN techniques. We also put detailed description on sample selection and data source. In chapter 3 we do the empirical study. We made comparisons between the MDA and ANN model. and find they do have very good predicting ability in financial distress for a period of four years. We find that ANN model is better than MDA model in predicting financial distress. In chapter 4 we summarize the empirical study results mentioned above. In addition, we summarize the shortcomings of our study and suggest the future study.