模糊贝叶斯网在通信行业客户流失预测中的应用研究

Application of fuzzy Bayesian net in customer churn prediction for mobile communication industry

文章针对移动通信公司客户的情况,为解决预测客户流失率时满意度模糊的问题,提出了通过使用语气算子、信息转换公式,使得模糊语言转化为精确概率,并将其整合到模糊贝叶斯网中,最终能够预测客户流失概率。实验证明,经过改进后的贝叶斯网,能够在一定程度上反映交流语言的模糊性,在信息模糊情况下能够较好地预测客户的流失情况。

In view of the fuzzy satisfaction problem in predicting the churn rate of Mobile Communications Company′s customers,the vague language is transformed into a precise probability through tone operator and information transformation formulas,and then integrated into the fuzzy Bayesian net,thus predicting the probability of customer churn.As practical results show,the improved Bayesian network reflects the ambiguity of communication language to a certain degree,and predicts the loss of customers more effectively in the ambiguous information situations.