吴祥佑. 基于驾驶行为的UBI车险定价模型[J]. 电子科技大学学报社科版, 2020, 22(4): 67-76. DOI: 10.14071/j.1008-8105(2019)-4014
引用本文: 吴祥佑. 基于驾驶行为的UBI车险定价模型[J]. 电子科技大学学报社科版, 2020, 22(4): 67-76. DOI: 10.14071/j.1008-8105(2019)-4014
WU Xiang-you. UBI Auto Insurance Pricing Model Based on Driving Behavior[J]. Journal of University of Electronic Science and Technology of China(SOCIAL SCIENCES EDITION), 2020, 22(4): 67-76. DOI: 10.14071/j.1008-8105(2019)-4014
Citation: WU Xiang-you. UBI Auto Insurance Pricing Model Based on Driving Behavior[J]. Journal of University of Electronic Science and Technology of China(SOCIAL SCIENCES EDITION), 2020, 22(4): 67-76. DOI: 10.14071/j.1008-8105(2019)-4014

基于驾驶行为的UBI车险定价模型

UBI Auto Insurance Pricing Model Based on Driving Behavior

  • 摘要:
    目的/意义随着5G时代的到来,更多地基于驾驶行为而非行驶里程定价将成为车险定价的关键。合理的UBI车险定价不仅有助于精算公平,更有利于UBI车险的普及推广。
    设计/方法基于UBI被保险人驾驶行为数据,构建了索赔次数影响因素的有序分类 logistic模型。
    结论/发现实证结果表明,高速行驶、急加速、违章次数和行驶里程数对出险索赔有显著的正向影响;急刹车和急转弯对出险索赔次数存在不显著的影响,其中急刹车将降低而急转弯则将提高出险次数。基于logistic模型所预测的被保险人在各索赔水平上的概率分布,可获得消除了运气扰动的期望索赔次数,以具体被保险人的期望索赔次数与全体被保险人平均索赔次数的比值为权重,可对车险实施较现行NCD定价更优的基于驾驶行为的定价。

     

    Abstract: Purpose/Significance With the advent of the 5G era, pricing based on driving behavior rather than mileage will become the key to auto insurance pricing. Reasonable UBI auto insurance pricing not only helps actuarial fairness, but also facilitates the popularization of UBI auto insurance. Design/Methodology Based on the driving behavior data of the UBI insured, an ordered logistic model of the influencing factors of the number of claims is constructed. Findings/Conclusions The empirical results show that high-speed driving, rapid acceleration, number of violations and mileage have a significant positive impact on insurance claims. Sudden braking and sharp turns have an insignificant effect on the number of insurance claims. Among them, sharp braking will decrease and sharp turns will increase the number of risks. Based on the probability distribution of the insured at each claim level predicted by the logistic model, the expected number of claims without luck disturbances can be obtained. Taking the ratio of the expected number of claims of the specific insured and the average number of claims of all the insured as the weight, it is possible to implement driving behavior-based pricing that is better than current NCD pricing for auto insurance.

     

/

返回文章
返回