基于支持向量机委员会机器的个人信用评估模型

Personal Credit Scoring Models based on Committee Machine of Support Vector Machine

为了充分利用SVM在个人信用评估方面的优点、克服其不足,提出了基于支持向量机委员会机器的个人信用评估模型.将模型与基于属性效用函数估计构造新学习样本方法结合起来进行个人信用评估;经实证分析及与SVM方法对比发现,模型具有更好、更快、更多适应性的预测分类能力.

In order to make full use of the strong points and to overcome the weak points of Support Vector Machine（SVM） on the credit scoring prediction problem, a personal credit scoring model is proposed based on committee machine of support vector machine（SVM- CM）. Utilizing this model together with the approach of using the utility functions estimated for attributes to extract learning samples from the credit scoring prediction problem, comparing its performance with SVM ,the experiment results show the model with better, quicker classification accuracy and being more compatible with classification problem.