基于随机森林算法的客户退保预警分析——以国联人寿为例

An Early Warning Analysis of Surrender Based on Random Forest Algorithm: Taking Guolian Life as an Example

面对日益激烈的市场竞争,对客户流失进行准确预警并提供有针对性的挽留服务,已成为寿险公司保证收入的重要因素。基于国联人寿保险公司数据管理现状,深度挖掘公司内部散落在系统中与退保相关的数据并形成原始因子,通过随机森林算法和阿里云机器学习平台对客户进行数据分析和建模。研究发现,利用公司内部已有零散客户信息,可构建出准确率较高的客户退保预警模型,帮助保险公司寻找客户退保原因并改进服务,为中小型寿险公司相关模型的建立打开新思路。

With the increasingly fierce market competition,early warning for customer loss and customer retention are important factors of revenue assurance for life insurance companies.Given the status of data management in Guolian Life Insurance Company,this paper explores the original factors using the surrender-related data that are scattered in the system.Random forest algorithm and Alibaba cloud machine learning platform are used for data analysis and modeling.This paper finds that an early warning model of surrender with high accuracy can be built by the customer information from the company.The model can help the company to identify the reason for surrender and improve service,which provides new ideas for modeling for small and medium-sized life insurance companies.