中文摘要：

在基于价值的客户细分中，不可避免地产生“拒真纳伪”的两类错误，由于错误分类代价差异和不同价值客户数量的不平衡分布，基于总体准确率的数据挖掘方法不能体现由于客户价值不同对分类效果带来的影响．本研究在代价敏感性学习机制下引入支持向量机作为分类工具，建立基于客户价值的错分代价函数，为适应客户价值多类别细分的要求，将二元分类扩展为多元分类，建立分类的期望损失函数作为分类效果的评价标准．实验结果说明，该方法可以更精确地控制代价敏感性和不同种分类错误的分布，降低总体的错误分类代价，使模型能更准确反映分类的代价，有效识别客户价值．

英文摘要：

Two type errors of ＂rejecting the true and accepting the false＂ are inevitable in customer value segmentation. The traditional data mining method which is based on the total accuracy rate can not reflect the influence caused by the great difference of misclassification costs and unbalanced quantity distribution of customers who have various values. The research proposes the cost-sensitive SVM classifer by presenting a misclassification cost function based on customer value, and expands binary-classification to multi- classification for the feature of customer segmentation, which is evaluated by the function of the exceptional lost. The data test result proves that the method can control the different types of error distribution with various cost of misclassification, reduce the total misclassification cost, and distinguish the customer value effectively.