Abstract
In recent years, China’s economy has undergone structural adjustments and a slowdown in growth. As a result, corporate credit risk has become increasingly important to monitor and assess accurately. Due to the low operational efficiency and high personnel costs associated with traditional credit risk assessment models, this study opted for an artificial intelligence model with support vector machines for risk assessment. To improve the accuracy and efficiency of the assessment model and make it more suitable for the financial characteristics of enterprises, a particle swarm algorithm was introduced to optimize the support vector machine model. For the choice of measurement indicators, the experiment extracted nine indicators with a cumulative contribution of 80% or more from 20 original indicators by means of principal component analysis as the experimental indicators. The model test results showed that the model had the greatest overall classification accuracy and recall rate, which were 89.44% and 58.83%, respectively. Meanwhile, the improved support vector machine model had the lowest misclassification rates for the first and second categories, at 24.45% and 16.74%, respectively. This indicates that the model is more likely to detect enterprises that can incur credit risk and can achieve the research objective with high applicability.
Get full access to this article
View all access options for this article.