摘要
在网络客户流失预测中,引入网络客户价值特征和情感特征,构建基于客户行为特征分析的网络客户流失预测模型结构。以国内某电子商务网站客户流失预测为实例,利用支持向量机与人工神经网络、逻辑斯蒂回归、朴素贝叶斯、决策树等方法进行对比研究,发现网络客户流失预测模型结构能在考虑预测模型准确性的条件下,兼顾预测模型的可解释性,是研究网络客户流失问题的有效方法。
A model structure on the analysis of customer behavior was constructed by using customer value feature and emotion feature for web customer churn prediction.The structure was used in web customer churn prediction for home electronic business carrier by comparing with support vector machine,artificial neural network,logistic regression,naive Bayesian classifier and decision tree.It is found that the structure has better accuracy with acceptable explanation of prediction model,and provides an effective approach for studying on two types of errors for web customer churn.
出处
《管理学报》
CSSCI
北大核心
2018年第3期442-449,共8页
Chinese Journal of Management
基金
国家自然科学基金资助项目(70801021)
广西高等学校高水平创新团队及卓越学者计划资助项目
广西跨境电商智能信息处理重点实验室培养基地专项资助项目
广西财经学院创新治理与知识产权学科群专项资助项目