C&R Tree算法在车险费率厘定中的应用

提出基于C&R Tree算法的车险理赔风险模型。现有研究方法对变量多、数据类型复杂和数据量大的数据不能准确进行分析,导致对车险客户的理赔风险出现误判,从而不能准确制定车险费率,为了解决上述问题,提出C&R Tree算法。C&R Tree算法通过检测输入字段,通过度量各个划分产生的异质性的减少程度,找到最佳的一个划分,将C&R Tree算法应用到车险理赔数据中,实验证明:所得理赔风险模型不依赖于经验知识,其模型易于理解,且具有较高的准确率,能够满足评价的要求。

Application of C＆R Tree Algorithms in Ratemaking of Auto Insurance Claims

In this paper,a risk model of auto insurance claims based on C＆R Tree Algorithm is put forward. For the current research methods,data with numerous variables,complex data types and large data volume can not be accurately analyzed,resulting in misjudgment of auto insurance customer claims so that auto insurance rates can not be developed accurately. To solve the above problems,C＆R Tree algorithm is proposed. C＆ RT algorithm can detect the input field and reduce the degree of heterogeneity generated by measuring individual partition to find the best partition. The algorithm is applied to auto insurance claims data,and experimental results show that the claims risk model does not rely on empirical knowledge,the model is easy to understand,and has high accuracy to meet the evaluation requirements.