【摘要】 提出了基于系统模糊优选和支持向量机的客户信用监测预警模型,该方法首先通过系统模糊优选模型划分样本类别,处理了不确定性信息,标示出样本的信用警级,其次利用标示样本对支持向量机进行训练,建立基于二叉树的多类支持向量预警模型;最后,对航运企业客户信用进行监测预警实验,并与BP神经网络和Logit模型进行对比,实验表明该方法有效、可行。更多还原

【Abstract】 A customer credit monitoring and early warning model supporting vector machine(SVM) based on system fuzzy optimum selection is presented.In this model,samples are classified by system fuzzy optimum selection and are marked with credit warning class.SVM is trained together with the classified sample.A multi-class support vector early-warning model based on binary tree is set up.The results of the comparison with BP and Logit models show that this model is effective and feasible.更多还原