基于模糊积分的企业信用评级方法研究

Study on A Credit-rating Approach for Corporation based on Fuzzy Integral

目前我国银行对于企业信用评级一般采用的是基于评分的等级分类方法,国内的研究虽然在方法上有大的改进,但是整体来看忽视信用评级过程的模糊本性。本文采用一种模糊信用评级方法,做一个有等级之分的决策结构的模型,运用模糊积分自下而上地综合信用信息。这种技术不仅考虑了目标数据,还考虑了每种指标的相关重要性。还有,这种方法使用了模糊集来描述指标,这样,最终的信用评级结果能揭示信用信息的变化。描述最终评价结果的五个等级水平可以提供更多有价值的信息。

Currently, our national banks usually use a hierarchical approach based on a point system in credit rating for commercial corporations. Although there are great improvements on rating approach in domestic study, the fuzzy nature of the credit-rating processes may be neglected. This paper presents a fuzzy credit-rating approach: first establish a hierarchical decision structure model and then employ fuzzy integral to aggregate the credit information in a bottom-up way. This technique considers not only the objective evidence but also the relative importance of each criterion. Furthermore, it uses fuzzy sets to describe the criteria, so that the final credit-rating results can reveal the changes of credit information. The five rating levels that describe the final evaluation results can provide more valuable information.