零膨胀模型在车险理赔次数中的应用与研究

发布时间：2018-01-25 09:47

本文关键词： 零膨胀模型机动车辆保险理赔次数费率厘定　出处：《黑龙江大学》2014年硕士论文　论文类型：学位论文

【摘要】：机动车辆保险,简称车险,是非寿险业务中占比最大的险种,其发展前景十分广阔.为了解决机动车辆保险的分类费率厘定的问题,当今保险公司经常使用的方法是广义线性模型.但当理赔次数出现大量的零赔付时,即出现零膨胀现象,广义线性模型将不再适用,在这种情形下,零膨胀回归模型可以解决此问题.一般我们假定零膨胀回归模型的结构零的比例参数是常数且不受车险费率因子的影响,但这有可能会和实际情况不符,因此本文假定结构零的比例参数不是常数且受其车险费率因子的影响. 本文主要介绍了机动车辆保险的传统理赔次数模型以及零膨胀模型,并且通过一组经典的保险数据进行分析,发现零膨胀模型能较好地处理理赔次数中大量零赔付的问题,模型中的参数估计更加精准,模型得出的结论也更切合实际情况,为从事保险行业的非寿险精算师们提供了一个解决此类问题的方法.
[Abstract]:Motor vehicle insurance, referred to as vehicle insurance, accounts for the largest proportion of non-life insurance business, and its development prospects are very broad. In order to solve the problem of classification rate determination of motor vehicle insurance. Nowadays, insurance companies often use the generalized linear model, but when there is a large number of zero claims, there will be zero expansion, the generalized linear model will no longer be applicable, in this case. The zero expansion regression model can solve this problem. Generally, we assume that the structure zero proportional parameter of the zero expansion regression model is constant and independent of the car insurance rate factor, but this may not be in accordance with the actual situation. Therefore, this paper assumes that the proportional parameter of structure zero is not constant and is affected by the car insurance rate factor. This paper mainly introduces the traditional claim number model and zero expansion model of motor vehicle insurance, and through a group of classic insurance data to analyze. It is found that the zero-expansion model can deal with the problem of a large number of zero claims in the number of claims, the parameter estimation in the model is more accurate, and the conclusion of the model is more practical. A solution to this problem is provided for non-life actuaries in the insurance industry.
【学位授予单位】：黑龙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F840

【参考文献】