基于贝叶斯网的分类器及其在CRM中的应用

Bayesian Network Classifier and Its Application in CRM

基于贝叶斯网的分类器因其对不确定性问题有较强的处理能力,因此在CRM客户建模中有其独特的优势。在对朴素贝叶斯分类器、通用贝叶斯网分类器优缺点分析的基础上,引入增强型BN分类器和贝叶斯多网分类器,详细介绍了后者的算法,并将其应用到实际电信CRM客户建模中,取得较好的效果。

Bayesian network classifier has good capability to handle problems with high uncertainty, and is fit for customer modeling in CRM. Based on the analysis of naive Bayesian classifier and general Bayesian network classifier, augmented naive Bayesian classifier and Bayesian multi-net classifier were introduced, and the detailed algorithm of the latter one was described. We applied Bayesian multi-net classifier in customer modeling of telecommunications CRM, and got effective results.