Abstract: Small enterprises credit risk evaluation is an issue of banks’ risk management, and is also related to the economic and social stability. For small enterprises loans, the default samples are far less than the non-default samples. Also, the impact on banks of default customers’ misjudgment is much larger than the impact on banks of non-default customers’ misjudgment. In this paper, this research will use the unbalanced support vector machine to weight the evaluation index and construct the credit scoring model for small enterprises. Firstly, the index weights are determined, according to the influence of the specific evaluation index on default and the influence of the index data on default. The greater the influence on default is,the bigger the weight will be.Secondly, using the correct rate of default samples and F-score as the evaluation model test standard, it will change the phenomenon that the overall accuracy is high, but the default samples’ accuracy is not high, which is caused by the unbalanced data. Finally, an empirical study of a Chinese national commercial bank’s of 3111 small enterprises loan observations is given and the result shows that industry climate index, capital immobilized ratio, net cash levels, Engel’s coefficient, operating profit are the key index of credit rating model of small enterprises.

Key words: credit rating evaluation, small enterprises’ loans, unbalanced support vector machines

摘要： 小企业信用风险评价既是银行风险管理问题,又事关经济社会稳定。针对小企业贷款实践中,违约样本远少于非违约样本、且违约客户误判对银行影响较大的现实,采用不均衡支持向量机对小企业信用风险评价指标进行赋权,进而构建了能有效区分违约客户、非违约客户的评价模型。根据有无特定评价指标、特定评价指标数值变化对贷款小企业违约状态的影响程度赋权;反映了对违约状态影响越大、评价指标权重越大的赋权思路。将违约样本正确识别率、违约样本的准确率与查全率等因素作为支持向量机赋权模型中客户识别率的度量标准,改变了样本数据不均衡所导致的样本总体精度很高、违约样本精度反而不高的现象。研究结果表明:行业景气指数、资本固定化比率、净利润现金含量、恩格尔系数、营业利润率等评价指标对小企业信用风险的影响较大。

关键词: 信用风险评价, 小企业贷款, 不均衡支持向量机