摘要
针对客户关系管理中的客户流失问题,建立了基于支持向量机的预测模型。基于实际客户流失数据样本数据量大、正负样本分布不平衡的特点,提出了一种改进支持向量机算法,并将其用于电信行业的客户流失预测。通过实际电信客户数据集测试,与传统的预测算法比较,证明这种算法适合解决大数据集和不平衡数据,具有更高的精确度。
To deal with customer churn problem in Customer Relationship Management (CRM) systems, prediction model based on Support Vector Machine (SVM) was set up, Due to large-scale and imbalanced churn data, an improved SVM- Imbalance Core vector machine SVM(ICSVM) was presented to predict customer churn, which has better arithmetic performance than others based on the test of real telecom data set. It was demonstrated that this algorithm was suitable for solving large-scale data set and imbalanced data with higher precision.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2007年第1期202-207,共6页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(70202008)。~~
关键词
客户流失
支持向量机
客户关系管理
预测
模式识别
customer churn
support vector machine
customer relationship management
prediction
pattern recognition