基于a股市场数据logistic回归的企业信用评级实证研究-金融工程专业论文.docxVIP

ABSTRACTABSTRACT ABSTRACT ABSTRACT This paper introduced several credit risk management methods and analyzed their advantages and disadvantages．The purpose of this paper is to find a credit risk rating model which can take the advantages of existing methodologies and performs well．This paper focused on the logistic regression model and KMV model，also analyzed their performances in the application of credit risk rating．Considering the logistic model has a stabile discrimination performance and KMV model has the ability of taking full use of capital market information，we added the variable of Distance to Default into the logistic model which was calculated from KMV model， and then we created a default prediction model which contained the dynamic market information．Then we selected default companies and corresponding good companies as training sample from Shanghai Stock Exchange based on the principle of “continuous profit turns to loss”．With the training sample，we created a new default prediction model and analyzed its performance．According to the ssme principle，we selected validation sample from Shenzhen Stock Exchange．With the validation sample，we verified the model to judge its accuracy and stability．Empirical studying shows that the variable of Distance to Default has a great help to improve the performance of the credit rating model．The logistic regression model adding the Distance to Default has a higher discrimination accuracy and stabil ity,which can be used as an ideal credit default prediction model． Key Words：Credit Risk，Logistic Regression，KMV Model，Distance to Default， Decision Tree，Default Prediction n 中国科学技术大学学位论文原创性声明本人声明所呈交的学位论文，是本人在导师指导下进行研究工作所取得的成 中国科学技术大学学位论文原创性声明 本人声明所呈交的学位论文，是本人在导师指导下进行研究工作所取得的成 果。除已特别加以标注和致谢的地方外，论文中不包含任何他人已经发表或撰写 过的研究成果。与我一同工作的同志对本研究所做的贡献均已在论文中作了明确 的说明。 作者签名： 翟：整里 签字日期： 坠!：3 中国科学技术大学学位论文授权使用声明 作为申请学位的条件之一，学位论文著作权拥有者授权中国科学技术大学拥 有学位论文的部分使用权，即：学校有权按有关规定向国家有关部门或机构送交 论文的复印件和电子版，允许论文被查阅和借阅，可以将学位论文编入有关数据 库进行检索，可以采用影印、缩印或扫描等复制手段保存、汇编学位论