ABSTRACT
The credit rating of research in China started late, and based on own limitations of SMES, such as less credit records, caused the financing diculties which severely curb the development pace of independent innovation. The basic way to solve this problem is to establish a set of high accuracy and specifically evaluation model for SMES. At present, the domestic development has also been gradually, Originally, the neural network technology was introduced to the commercial bank credit risk assessment in [1], and then gradually the decision tree, genetic algorithm and support vector machine (SVM) method were also applied to risk assessment, now many scholars made many improvements on them, but nobody compare these methods .In this paper, we utilize four classical methods of machine learning, including Support Vector Machine (SVM), Decision Tree, Random Forest and Boosting, to model the risk assessment for SMES and make a comparison to select the most appropriate model. The contribution of the paper is listed as follow:
1 Many scholars introduced dierent methods of machine learning into credit evaluation, however, no person make a comparison to these ways and select the most appropriate method to model credit assessment.
