基于粗糙集理论的信贷风险评估模型研究

Research on Credit Risk Assessment Model Based on Rough Set Theory

信贷风险是现代商业银行所需面对的首要风险,特别是我国商业银行由于缺乏基础数据,无法采用国外银行的先进信贷风险评估模型,长期以来一直使用传统方法进行信贷风险评估,因此急需探索一个适用于我国国情的信贷风险评估模型。为此,首先建立一套包含财务指标与非财务指标的信贷风险评估指标体系,然后根据粗糙集理论能够处理不可区分关系的特点,结合我国具体国情,提出了基于粗糙集理论的信贷风险评估模型,并给出数据预处理、属性简化、决策规则集的生成、对象分类及规则预测精度验证的实现方法。最后以多家公司的信贷情况为测试实例,采用基于粗糙集理论的信贷风险评估模型进行测试,测试结果表明,信贷正常公司的预测准确率达到83.33%,非正常公司的预测准确率达到100%,能够为银行的信贷决策提供有效的参考。

Credit risk is the primary risk for which modern commercial banks are lacing, especially in our country due to lack of basic data, so commercial banks can＇t use the tbreign advanced credit risk assessment model to have to use the traditional one so as to explore a suitable tbr China＇s national conditions of the credit risk assessment model. Aiming at this problem, first of all, a set of financial indicators and non-financial indicators of credit risk assessment index system should be established and then according to the characteristic, which theory ot＂ rough set is able to handle indistinguishable relationship, combining with China＇s specific national conditions, the credit risk assessment model based on rough set theory is put tbrward and presents a simplified data preprocessing, attribute and decision rule set the generation rules, object classification and prediction accuracy of the implementation of the method. Finally the multiple companies credit conditions are tested for some cases with the credit risk assessment model based on rough set theory. The results show that the prediction accuracy in credit normal companies reaches 83.33%, it is 100% in abnormal companies. Which could provide an effective reference for bank credit decisions.