汽车保险奖惩系统及其应用研究
发布时间：2019-02-24 10:29
【摘要】：汽车保险奖惩系统是应用于续期保费阶段的一种重要经验估费方法,其主要工作原理是依据历史索赔信息对投保人进行保费“区别”对待:针对在上个保险年度内没有发生索赔事件的投保人实行减免续期保费的“奖励”,而对发生过一次或多次索赔要求的投保人在相应程度上给予增收保费的“惩罚”。本文主要围绕基于索赔次数模型的奖惩系统及其工作原理展开研究讨论,具体内容和研究成果如下:1.鉴于保险实践中无索赔次数记录出现的保单通常在保单组合内占有较大比重,本文采用叠加和调零后的复合分布类来拟合同质性保单的索赔次数,并分别给出了参数估计的具体方法;2.奖惩系统等级转移概率均由其在上一保险年度内的索赔次数所决定,利用这一工作原理,通过时间序列方法中的INAR(1)模型对下一保险年度的索赔次数进行了预报估计;3.介绍了经典奖惩系统模型的基本数学理论和保费选择方法,从建立保费等级,讨论等级转移共性法则,计算索赔发生概率并得到概率转移矩阵,到最优奖惩系统,不同保费原理的选择及保费原理下经验费率系数的确定等;4.考虑到只依据索赔次数分布构建的奖惩系统自身存在的局限性,在真实索赔次数模型的基础上,通过设立标准索赔额基数对索赔金额进行分类,构建出一个新随机变量——高额索赔次数,同时完成高额索赔次数的分布模型假设及参数估计;5.针对我国保险公司在车险经验定价过程中所遇到的实际问题,提出了对现行奖惩制度的三大改进建议。包括将投保人违章记录纳入车险奖惩系统;构建路径依赖的奖惩系统;以及利用互联网信息技术搭建信息共享平台,从而预防高风险投保人出现逃避缴纳溢价保费的行为。
[Abstract]:The automobile insurance reward and punishment system is an important empirical method used in the renewal premium stage. Its main working principle is to treat the policyholder's premium "differently" according to the historical claim information: "reward" for the policy holder who did not have a claim event in the last insurance year. Policy-holders who have made one or more claims are punished to a certain extent. This paper mainly focuses on the reward and punishment system based on the number of claims model and its working principle. The specific contents and research results are as follows: 1. In view of the fact that the number of claims recorded in the insurance practice usually occupies a large proportion in the policy portfolio, this paper uses the composite distribution class after superposition and zero adjustment to fit the claim number of the homogeneous policy. The methods of parameter estimation are given respectively. 2. The probability of grade transfer of reward and punishment system is determined by the number of claims in the last insurance year. Using this working principle, the number of claims in the next insurance year is forecasted by the INAR (1) model in the time series method; 3. This paper introduces the basic mathematical theory of the classical reward and punishment system model and the method of premium selection, from establishing the premium grade, discussing the general law of grade transfer, calculating the probability of claim occurrence and obtaining the probability transfer matrix, to the optimal reward and punishment system. The choice of different premium principle and the determination of experience rate coefficient under premium principle; 4. Considering the limitations of the reward and punishment system based only on the distribution of the number of claims, and on the basis of the true number of claims model, the amount of the claim is classified by establishing a standard base of claim amount, A new random variable, the high number of claims, is constructed, and the distribution model hypothesis and parameter estimation of the high number of claims are completed at the same time. 5. In view of the practical problems encountered by Chinese insurance companies in the process of car insurance experience pricing, three suggestions for improving the current reward and punishment system are put forward. Including the policy holder violation records into the vehicle insurance reward and punishment system; build a path dependent reward and punishment system; and use Internet information technology to build information sharing platform to prevent high-risk policyholders from paying premium premiums.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F842.634
本文编号：2429463
[Abstract]:The automobile insurance reward and punishment system is an important empirical method used in the renewal premium stage. Its main working principle is to treat the policyholder's premium "differently" according to the historical claim information: "reward" for the policy holder who did not have a claim event in the last insurance year. Policy-holders who have made one or more claims are punished to a certain extent. This paper mainly focuses on the reward and punishment system based on the number of claims model and its working principle. The specific contents and research results are as follows: 1. In view of the fact that the number of claims recorded in the insurance practice usually occupies a large proportion in the policy portfolio, this paper uses the composite distribution class after superposition and zero adjustment to fit the claim number of the homogeneous policy. The methods of parameter estimation are given respectively. 2. The probability of grade transfer of reward and punishment system is determined by the number of claims in the last insurance year. Using this working principle, the number of claims in the next insurance year is forecasted by the INAR (1) model in the time series method; 3. This paper introduces the basic mathematical theory of the classical reward and punishment system model and the method of premium selection, from establishing the premium grade, discussing the general law of grade transfer, calculating the probability of claim occurrence and obtaining the probability transfer matrix, to the optimal reward and punishment system. The choice of different premium principle and the determination of experience rate coefficient under premium principle; 4. Considering the limitations of the reward and punishment system based only on the distribution of the number of claims, and on the basis of the true number of claims model, the amount of the claim is classified by establishing a standard base of claim amount, A new random variable, the high number of claims, is constructed, and the distribution model hypothesis and parameter estimation of the high number of claims are completed at the same time. 5. In view of the practical problems encountered by Chinese insurance companies in the process of car insurance experience pricing, three suggestions for improving the current reward and punishment system are put forward. Including the policy holder violation records into the vehicle insurance reward and punishment system; build a path dependent reward and punishment system; and use Internet information technology to build information sharing platform to prevent high-risk policyholders from paying premium premiums.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F842.634
【参考文献】
相关期刊论文 前2条
1 夏冬;肖宇谷;孟生旺;;中国现行汽车保险奖惩系统的评价[J];统计与决策;2008年15期
2 孟生旺,袁卫;汽车保险的精算模型及其应用[J];数理统计与管理;2001年03期
相关硕士学位论文 前5条
1 袁小康;我国车险奖惩系统的实证研究[D];西南财经大学;2011年
2 朱建清;机动车辆保险BMS模型的设计[D];山东科技大学;2009年
3 高玮;对汽车保险奖惩系统的研究[D];北京交通大学;2008年
4 吴建祥;车险奖惩系统及其定价[D];西安理工大学;2007年
5 付海燕;汽车保险的奖惩系统分析[D];华东师范大学;2006年
，本文编号：2429463
本文链接：https://www.wllwen.com/jingjilunwen/bxjjlw/2429463.html
