Enterprise Credit Rating Method Based on Scalable SVM Decision Tree Algorithm

Abstract

With the growing number of enterprises and the increasing complexity of economic activities, the significance of enterprise credit rating has become more pronounced. Traditional credit rating methods often face issues such as lack of precision and limited scalability, making them inadequate for the modern economic environment. To tackle these challenges, this paper presents an enterprise credit rating method utilizing a scalable support vector machine (SVM) decision tree algorithm. By integrating the advantages of SVM and decision trees, this approach improves rating accuracy through the construction of scalable SVM classifiers within a decision tree framework. Experiments conducted on the Shenzhen manufacturing enterprise dataset spanning 2018 to 2023, along with comparative analyses against traditional logistic regression, random forest, and gradient boosting decision tree methods, validate the superior performance and accuracy of the proposed approach. Our model achieves 90.8% in precision, 86.7% in recall, and 0.935 in AUC. Future research will aim to further optimize the algorithm's complexity and extend its application to other fields.