摘要
为了实现供应链金融信用风险的科学定量管理,建立了一种AdaBoost集成的支持向量机(SVM)算法。该方法首先采用SVM方法对信用风险数据进行分类学习,建立基学习器;接着通过AdaBoost集成算法对基学习器迭代训练,生成最终的供应链金融信用风险评估模型。实证结果表明,AdaBoost集成SVM分类器较模糊积分SVM集成等方法具有更高的分类准确率,因此该模型具有很好的应用前景。
AdaBoost-integrating support vector machine （SVM） classifier is established in order to realize scientific measurement management of supply chain finance credit risk in this paper. The credit risk data is classified and learnt by firstly using SVM method with an establishment of base classifier. Then, a final evaluation model for sup- ply chain finance credit risk is built through iterative training of AdaBoost algorithm. The empirieal results show that the AdaBoost integrated SVM classifier has higher classification accuracy than that using fuzzy integrated method, therefore this model has a bright application prospect.
出处
《征信》
北大核心
2014年第11期19-22,共4页
Credit Reference
基金
上海市教委科研创新项目(12YZ1710)
上海金融学院校级科研项目(SHFUKT13-07)
