具有通胀因子的平衡分位回归信度模型

浏览次数： 11

分享到:

摘要:

保险数据经常表现为尖峰或厚尾的形态,而经典的信度模型并不能有效反映分布的上尾和下尾信息,同时考虑保费的安全负荷性以及通货膨胀对未来保险期间对保额的影响.在分位数视角下用平衡损失代替均方损失,通过利用样本分位数充分利用所有数据的信息,建立了具有通货膨胀因子的分位数信度模型,扩展了经典的信度模型,为保险公司提供了一种根据实际情况选择不同的权重,制定更符合实际保费的方法.

Insurance data often appear as spikes or thick tails,but the classical reliability model does not effectively reflect the distribution of upper and lower-end information.Consider the safety load of premiums and the impact of inflation on the amount of coverage for future insurance periods.From the quantile perspective,the mean square loss is replaced by the balanced loss,by using the sample quantile to make full use of the information of all data,a quantile reliability model with inflation factor is established.Moreover,it extends the classical reliability model and provides insurance companies with different weights according to the actual situation.Finally,a more realistic premium approach can be developed.

作者:

程纪王金瑞吴黎军

Cheng Ji;Wang Jinrui;Wu Lijun(College of Mathematics and System Sciences,Xinjiang University,Urumqi 830046,China)

机构地区:

新疆大学数学与系统科学学院

出处:

《betway官方app 学报：自然科学版》 CAS 北大核心 2020年第4期1-6,共6页

基金:

国家自然科学基金(11361058 11861064)。

关键词:

平衡损失函数通货膨胀分位数回归信度模型

balance loss function inflation quantile regression credibility model