不完全信息下的一种信用分类方法

A Credit Classification Method for Incomplete Information

本文详细分析了一类典型的不完全信息的信用评价问题,即存在信用参考信息不完全或样本数据较少,以及某些指标数据为区间数等。文章首先考虑具有不完全信用参考信息的情况,在分析这类问题特性的基础上,基于数据包络分析理论(DEA),提出一种以拒绝案例集构造参考单元集的方法,并给出了相应的DEA模型。然后,进一步考虑带有较为复杂区间指标数据这类具有不完全指标信息的信用评价问题,并给出了相应的信用分析方法,最后采用算例说明本文提出方法的合理性。由于所提方法能有效地处理信用信息不完全和指标数据不完全的情况,故其具有较为广泛的实用性。

This paper analyzes a typical credit evaluation problem with incomplete information, such as in- complete credit reference information and interval data. After analyzing the problem with incomplete credit reference information and the characteristics of the problem, one method based on data envelopment analysis （DEA） theory is presented to set a new reference cells with the collection of rejection cases, and a corresponding DEA model is proposed. Moreover, considering credit evaluation measures in the presence of complicated interval index value, a credit classification method is shown. And a numeric example is em- ployed to illustrate the method. Since the present method can effectively solve the credit evaluation problem with incomplete credit reference information and interval data, it can be easily integrated into many practical applications.