Enterprise credit risk assessment method, apparatus and device, and storage medium

Abstract

The invention discloses an enterprise credit risk assessment method. According to the method, enterprise tax data is received, the operation credit risk of an enterprise is quantified from the dimension of the tax data, and a foundation is laid for the precise risk assessment of enterprise credit; meanwhile, a small and micro enterprise credit risk model called by the method is built based on an XGBoost algorithm, and the feature cross ability of weak variables of the model is guaranteed; based on tax administration sample data analysis in the training process, variable stability of sample variable data and model stability serve as evaluation indexes after variables are preprocessed, analog input variable screening is conducted on the sample variable data, the influence of abnormal samplevariables on model training can be filtered, the problem of overfitting when the XGBoost algorithm is adopted for the small and micro enterprise credit risk model is relieved, and the enterprise credit risk assessment accuracy is improved. The invention further provides an enterprise credit risk assessment apparatus and device and a readable storage medium, which have the above beneficial effects.

Description

Technical field [0001] This application relates to the technical field of anterior segment inspection, and in particular to an enterprise credit risk assessment method, device, equipment and a readable storage medium. Background technique [0002] The widespread application of big data and Internet technology has a profound impact on my country’s financial ecology and at the same time provides new platforms and channels for small and micro enterprises to raise funds; the innovative application of big data technology in the field of Internet finance is more important for small and micro enterprises The development of financial business has created more possibilities. [0003] With the development of these technologies, corporate credit reporting has also developed. The corporate credit reporting system can solve some information asymmetry problems, reduce information costs and transaction costs, and thereby reduce adverse selection. The credit information system can collect, process...

AI Technical Summary

Problems solved by technology

[0004] At present, the traditional logistic regression model is commonly used in credit risk models. Although logistic regression has relatively good business interpretability, it cannot learn the characteristic cross-ability model of some weak variables in the Internet era. Therefore, more and more Machine learning algorithm applied to the credit risk model of small and micro enterprises

[0005] At present, the XGBoost method is often used for data processing in credit risk models of small and micro enterprises. XGBoost (eXtreme Gradient Boosting) is an integrated learning data processing method. The method of data processing will cause the model to be prone to overfitting. After the model is overfitted, the generalization ability of the model will be weak, which will affect the recognition accuracy of the model.

Images