论文部分内容阅读
保险索赔历史数据不足,可靠性不高,将严重影响保险公司对未决赔款准备金的评估精度。本文借鉴国内外相关研究经验,研究了准备金评估模型中的平滑问题,介绍了一种以广义线性模型为框架,考虑事故年因子、发展期因子和日历年因子为主要影响因素的平滑化模型,并结合利用Bootstrap法,给出了模型的参数估计方法和未决赔款准备金的预测分布与均方误差等其他统计特征度量。平滑化方法可有效减少随机模型的参数个数,使得参数估计变平稳,提高模型的索赔尾部数据处理能力。
Insufficient historical claims of insurance claims and low reliability will seriously affect the accuracy of insurance companies’ assessment of the outstanding compensation reserves. This paper studies the smoothing problem in the reserve assessment model by referring to the relevant research experiences both at home and abroad. The paper introduces a smoothing model which takes the generalized linear model as the framework and the factor of the accident, the development factor and the calendar year as the main factors , And combined with the use of Bootstrap method, given the parameters of the model estimation methods and predictive distribution of residual claims reserves and mean square error and other statistical features measure. The smoothing method can effectively reduce the number of parameters of the stochastic model, stabilize the parameter estimation and improve the data processing ability of the tail of the model claims.