基于GP的个人信用评估非线性组合预测模型

Personal Credit Scoring Model of Non-linear Combining Forecast Based on GP

个人信用评估对于商业银行控制信贷风险具有重要意义。针对单一模型存在的分类精度不高的问题,需将组合预测模型用于个人信用评估。本文在线性回归和Logistic回归两种单一统计模型的基础上,利用遗传规划(GP)构建了一种非线性组合预测模型。将模型应用于某商业银行的消费信贷数据的分类,其结果表明,基于GP的非线性组合预测模型有效地提高了分类精度,模型的第二类误判率低,对于商业银行控制个人信用风险具有更好的适用性。

Personal credit scoring plays an important role for commercial banks to control consumer credit risks. Aiming at the low predictive accuracies of single models, this paper presents a combining forecast model for personal credit scoring. Based on two single statistical models of linear regression and logistic regression, this paper constructs a non-linear combining forecast based on genetic programming （GP） and uses the constructed model to classify the consumer credit data of one commercial bank. The application results indicate that the non-linear combining forecast based on GP increases the predictive accuracy effectively and the model also gets a much lower type Ⅱ error rate which is more applicable for commercial banks to control consumer credit risks.