基于SVM算法的客户诚信度评价模型的研究

Research on Customer Credit Evaluation Model Based on SVM Algorithm

针对目前大量的商业活动中对客户诚信度评价方法的欠缺,提出了一种新型的预测模型,即将高斯核函数同SVM算法相结合,在非线性多维特征解空间的环境下,也可做到较为准确的评估.通过实践验证,同以往的预测模型相比较,该算法不论在预测准确性以及泛化能力上都有更为突出的表现.

To solve the lacks of customer credit evaluation method which general used in commercial activities, This paper proposes a new prediction model which combined the Gauss kernel function with SVM algorithm. This model could get more accurate assessment in the environment of nonlinear multidimensional solution space. Compared with the previous prediction model, the algorithm has more prominent performance prediction accuracy and generalization ability