Prediction and Analysis of Corporate Bond Risk Based on Data Mining

Abstract

The advent of the era of big data and the development of the Internet and computer technology have brought new tools for risk early warning to the bond market. From the perspective of micro risk information of the company’s internal operation, this paper uses data mining technology to find out the key factors affecting the maturity of corporate bonds, and establishes a method to predict bond default. Using the extreme gradient boosting algorithm, it is found that the main influencing factors of whether the bond defaults are the year-on-year growth rate of operating revenue and asset liability ratio in this paper. And then we establish a binary logistic regression model of the bond. The binary logistic regression model can predict whether the bond defaults.