ABSTRACT
The importance of enterprise credit risk assessment is growing, and with the advent of digital economy, the explosion of social information has also spawned many credit data. Therefore, this paper evaluates the credit risk of enterprises based on the integrity behaviour and credit ability data of 292 enterprises. Based on the idea of data-driven, this paper evaluates the overall enterprise credit environment score from the macro and micro perspectives for the credit environment data. Choose an adaptive weighting method to evaluate the integrity behaviour of enterprises, and combined with improved support vector machine, data envelopment analysis and other methods to explore the fac-tors that affect the credit ability of enterprises. The k-nearest neighbor support vector machine model has high stability and prediction ability for enterprise credit ability; Enterprises can be divided into four categories according to their integrity behaviour and credit ability.
