Prediction of car insurance claim frequency is a focus of theoretical and empirical research of non-life actuarial studies.However, owing to the high-dimensional information involved, traditional models and estimation methods no longer apply.In this paper, some significant factors of car insurance claim frequency are identified through the variable selection method with convex penalty function based on the mean count model.A small simulation and a real data analysis are conduc-ted to assess the feasibility of the proposed model and methods.%汽车保险的索赔频数预测问题是非寿险精算理论和应用研究的一个重要内容。但是，在含有高维附加信息的情形下，传统的估计方法就不再适用。本文在均值计数模型基础上，利用凸惩罚函数进行变量选择，找到影响车险索赔频数的显著性因子，并通过模拟和实例分析来评价该模型和所提出的方法的可行性。
展开▼
