基于操作时间来重新设计流量三角形,通过对Hoerl曲线进行推广来刻画异质的损失进展模式,分别建立双广义线性混合模型和Tweedie分布簇广义线性混合模型来评估准备金。该模型可以综合信息平台汇总数据,同时利用个体经验数据和行业数据来评估任何保险公司在任何时点的准备金,解决传统准备金评估技术的弊端,为动态风险监管提供决策依据。在实证分析中,根据重疾险的数据特征,设计适合的准备金评估模型,最后进行对比分析,总结模型方法的优点、缺点和主要结论。
This article used operational time to redesign the runoff triangle, and constructed double generalized lin ear mixed models and Tweedie-based generalized linear mixed models by way of improving Horel curve to describe heterogenerious claim developments factors. While these models can integrate all-round statistics on the information platform, they can evaluate claims reserves at any time for any insurers based on their individual experiences and in- dustry data. Thus, they can address the deficiency of traditional reserve evaluation techniques, and can provide with decision-making references for dynamic risk management. In the exponential analysis, we constructed a claims reser ving model for short-term critical illness insurance, made a comparative analysis of the results, and summarized the strengths and weakness of the models and our main conclusions.