Empirical Research on Credit Evaluation of Real Estate Industry Based on Multi-source Data Fusion

Abstract

In this paper, based on the evaluation of corporate credit risk using traditional financial indicators, we will combine with non-financial data, define whether the listed company has ST in its abbreviation as a default fact, and establish a logistic model to quantitatively analyze the credit risk. In this paper, 123 listed companies in the real estate industry are used as samples, and the financial data, number of employees, percentage of shareholding of the largest shareholder, and CEO's education disclosed in the 2019 and 2020 annual reports are obtained through the Guotaian and CNRDS databases. Using principal component analysis to extract principal component factors from 21 financial data and combining non-financial indicators, a logistic model based on financial indicators and a logistic model based on financial and non-financial data were constructed to study the impact of non-financial data on company credit risk. It is found that non-financial indicators reflecting company size as well as corporate governance ability have a certain degree of influence on the credit risk of the company, i.e., the prediction results of the logistic model based on financial and non-financial data are more effective.