基于RFM模型的随机森林算法对民航客户的流失分析

Analysis of Airline Customer Churn by Random Forest Algorithm Based on RFM Model

近几年,随着航空市场的快速发展,对于航空公司而言,如何在增加市场占有率的同时,对客户的流失进行有效的控制也刻不容缓。基于随机森林算法,根据航空客户数据,建立流失预测模型,对客户是否已流失进行预测研究,将传统的RFM客户价值模型进行改进,结合随机森林算法对客户流失进行预测。实验结果表明,基于RFM模型的随机森林算法构建的客户流失模型拥有更具有说服力的指标选取,AUC值达到0.92,且准确率较高。利用该模型可对航空公司客户流失进行较为准确的预测,对流失客户进行分类,为民航企业提供营销策略。

In recent years,with the rapid development of the aviation market,it is urgent for airlines to control the loss of customers while increasing their market share.Based on the random forest algorithm,according to the data of aviation customers,a loss prediction model is established to predict whether customers have lost.The traditional RFM customer value model is improved and the random forest algorithm is used to predict customer churn.The experimental results show that the customer churn model based on RFM stochastic forest algorithm has a more persuasive index selection,an AUC value reaches 0.92 and the accuracy is higer.The model can be used to predict the loss of airline customers accurately,classify the lost customers and provide marketing strategies for civil aviation enterprises.