基于尺度混合偏正态分布的稳健未决赔款准备金评估方法

Robust Outstanding Claims Reserve Assessment Methods Based on Scale Mixtures of Skew-normal Distribution

赔款准备金评估作为保险精算的一个重要研究领域,对于含有异常值的赔款数据,很多学者不断探索更加合理的赔款准备金评估方法,本文应用稳健统计方法,即假设赔款准备金模型中误差分布服从尺度混合偏正态分布构成的厚尾分布,降低异常值的影响,得到更加合理的评估结果。通过实际数据分析,并与已有的稳健赔款准备金统计方法进行比较,基于尺度混合偏正态分布的稳健赔款准备金评估方法取得比较好的评估结果,具有实际应用价值。

Loss reserve assessment method is an important research field of insurance and actuarial science,many scholars continue to explore more reasonable method to deal with data containing abnormal claims.In this paper,we use the robust statistical method,that is,to assume that the error distribution in the loss reserve model has thick tail distribution of scale mixtures of skew-normal distribution,reduce the impact of abnormal values,to obtain more reasonable assessment results.Through the actual data analysis,and compared with the existing statistical methods,this paper has made a good evaluation result and has practical application value.