基于改进的AdaBoost算法的信用评分模型

Credit Score Model Based on Improved AdaBoost Algorithm

将AdaBoost组合算法应用于信用评分模型中的分类问题,并针对该算法在解决不平衡分类问题上的一些不足,对算法进行了改进。应用此改进的AdaBoost算法,创建了新的信用评分模型,并进行了实证分析。实证结果表明,基于改进的AdaBoost算法的信用评分模型可以有效降低由于模型错判而导致的损失。

This paper applies AdaBoost algorithm to classification problem in credit score models,and improves AdaBoost in consideration of the shortcoming emerged when dealing with unbalanced data.This improved AdaBoost is used to build credit score models.Empirical results show that credit score models based on improved AdaBoost algorithm can decrease the loss caused by misclassification.