Default Risk Assessment of Internet Financial Enterprises Based on Graph Neural Network

Abstract

In recent years, the increasing number of default events happened in the Internet financial enterprises has incurred great financial losses to investors. Early warning to the enterprises with high default risk is of great significance to protecting the benefit of investors. There exist two challenges in traditional default risk assessment methods: poor data availability and neglect of risks from the affiliated entities. In order to address these problems, we collect the Internet financial enterprises’ nonfinancial data and propose a default risk assessment method for Internet financial enterprises based on heterogeneous graph neural network. Extensive experiments on a real-world dataset show that the proposed method outperforms the baseline models on the task of default risk assessment.