粗糙集属性约简在客户流失预测中的应用

Application of Rough Set Attribute Reduction in Customer Churn Prediction

客户流失预测作为客户关系管理的主要问题,一直受到研究学者们的关注。企业通过内部和外部的数据信息,对客户流失情况进行预测,针对还未流失但有流失倾向的客户采取相应的营销策略。大数据时代使得数据信息爆炸式增多,如何处理高维数据信息成为客户流失预测的难点。利用粗糙集理论进行属性约简可以降低数据维度,并有效地实现客户特征选择,从而降低客户流失预测的运算难度,提高预测性能。

Customer churn prediction,as the main problem of customer relationship management,has always been the focus of researchers.Based on internal and external data information,enterprises can forecast customer turnover and adopt corresponding marketing strategies for customers who have not yet lost but have a tendency to lose.The era of big data makes data information increase explodes,and how to deal with high-dimensional data information becomes the difficulty of customer churn prediction.The attribute reduction using rough set theory can reduce the data dimension and effectively realize the customer feature selection,so as to reduce the operational difficulty of customer churn prediction and improve the prediction performance.