CART决策树在旅游保险数据中的研究

Research on CART Decision Tree in Tourism Insurance Data

通过CART分类决策树,找到各指标对是否购买该公司旅游保险的影响程度,使得旅游保险行业更容易寻求到更多的潜在的在保人。对Tour&Traves公司提供的数千条客户基本信息进行处理,利用CART决策树算法对该数据进行分析以及建立模型,并对决策树模型进行评价和可视化输出。为了避免原模型过拟合的问题,在原有CART分类决策树模型进行后剪枝,在不减少模型正确率的情况下降低了模型的复杂度。

Through CART classification decision tree,the influence of each index on whether to purchase the company's travel insurance is found,which makes it easier for the travel insurance industry to seek more potential insurers.Thousands of basic customer information provided by Tour&Traves are processed,and the CART decision tree algorithm is used to analyze the data and establish a model,and the decision tree model is evaluated and visualized.In order to avoid the problem of over-fitting of the original model,the original CART classification decision tree model was pruned after the pruning,which reduced the complexity of the model without reducing the accuracy of the model.