基于KMOD核函数的SVM方法在信用评分中的应用

APPLICATION OF SVM BASED ON KMOD FUNCTION IN CREDIT SCORING

本文介绍了支持向量分类机,并引入具有更好识别能力的KMOD核函数建立了SVM信用卡分类模型.利用澳大利亚和德国的信用卡数据进行了数值实验,结果表明该模型在分类准确率、支持向量方面优于基于RBF的SVM模型.

In this paper, we give a desvription of support vector classification machine, and established support vector machine credit classification based on a better discrimination kernel function named KMOD. Experiment with the credit card data of Australia and German showed our model outperformed RBF based SVM in classification accuracy and support vectors.