Quantitative Analyses on the Default Rate of Non - listed SMEs:An Analysis based on the Discriminant Approach and the Decision Tree Model
用判别分析和决策树模型预测非上市中小企业违约率，发现两者效果均好，但决策树还能找出关键因素，如现金流/总债务比和流动资产/流动负债比，对银行审核信用有帮助。
This paper applies discriminant analytical approach and the Decision Tree Model to analyze the default rate of non-listed small and medium-sized enterprises(SMEs),and also compares the results generated from both approaches.The results show that both Decision Tree Model and Discriminant Analysis are performing well in predicting default probabilities.However in comparison with the Discriminant Analysis,Decision Tree Model has advantages in which it can not only predict default probabilities well,but also help us find out the critical factors that affect the firm's default rate.From the samples it is found that cash flow/total debt ratio and liquid asset/liquid debt ratio are two critical factors for commercial banks to verify credit status of the SMEs.It would greatly improve the accuracies of predicting default rate if banks can ensure those ratios are correct by ways of careful verification and investigation.