Scholar Publishing Group The best way to conference proceedings by Scholar Publishing Group

CONFERENCES
Scholar Publishing Group

Research on Credit Risk of enterprises based on Logistic Regression and Boost

Download as PDF

DOI: 10.38007/Proceedings.0001190

Author(s)

Ye Yang

Corresponding Author

Ye Yang

Abstract

In recent years, the number and amount of defaulted debentures in the bond market keep rising, causing annual economic loss more than 100 billion yuan and this trend is growing. The bond market is only a small part of the huge economic market. Because of the contagion effect, the credit risk of the entire market cannot be ignored. As a financing institution, one of the risks that banks faced is credit risk. The main method currently used by banks to measure credit risk is internal ratings based approach, which is a method with a large degree of freedom. AS the global trend tighter regulation deepens, because of the drawbacks of excessive freedom of internal ratings based approach, Basel III restricts it.Therefore, looking for a new method to measure enterprises’ credit risk has become a consideration for financial institutions. This paper tries to find a new method to measure credit risk by using machine learning method.

Keywords

Credit Risk; Logistic Regression; XGBoost