摘 要

我国保险公司为适应车险费率市场化，需要推出更加科学的车险确定机制。由于车辆保险风险因子种类和数量较少，保险产品单一，无法体现出驾驶人之间的差异性和个性化特点。因此如何实现车险定价机制科学合理，让不同驾驶人承担更加精确合适的保费就成为车险市场化进程中亟待解决的问题。本文将基于大数据思维，通过所学专业的相关知识，研究如何将驾驶行为风险等纳入车险定价的方法。

首先，在分析传统车险的基础上，提出一种基于驾驶行为的费率厘定新思路，从而优化“从车”和“从人”两种费率厘定模式。传统车险的费率厘定从两个方面考虑，一是“从 车”模式，所考虑的风险因子包括车辆类型、车辆用途以及车龄等；二是“从人”模式，所考虑的风险因子包括驾驶人年龄、性别以及职业等。通过对车辆保险理赔成因分析，发现这两种模式考虑的风险因子均为静态因子，而实际交通事故发生的概率跟驾驶行为这一动态因子密切相关。因此车险费率厘定应充分考虑驾驶行为这一动态因子。

其次，基于大数据概念构建车辆安全相关因素风险评估模型。根据风险因子的选取原则，得到影响驾驶安全的主要风险因子。根据层次分析法的指标体系构建思想，建立风险评估体系。再通过熵权层次分析法对每个风险因子求取指标权重，确定每种风险因子的风险评估分数，建立风险评估模型。

最后，将相应的风险评估分数纳入到车险定价的风险因子中，在原有的费率调整系数上提出各风险因子系数，使得多因素风险能够反映到车险费率和车险保费上，并运用软件和实例分析对基于大数据的车险定价机制进行验证，证明基于多因素考虑的车险费率厘定方法的科学性。

本文所述基于大数据的多因素风险评估模型和车险费率厘定方法研究，对车险费率的科学性、合理性和公正性具有理论指导和实践应用价值，不仅能够推动车险行业的健康发展，而且对驾驶人驾驶行为有着纠正、提醒作用，促进人们文明驾驶。

关键词：大数据；汽车保险；汽车保险费率

Abstract

In order to adapt to the marketization of vehicle insurance rates, insurance companies in China need to introduce a more scientific vehicle insurance determination mechanism. Due to the fewer types and quantities of vehicle insurance risk factors, the single insurance products can not reflect the differences and personalized characteristics between drivers. Therefore, how to make the pricing mechanism of automobile insurance scientific and reasonable and let different drivers bear more accurate and appropriate premium has become an urgent problem to be solved in the process of automobile insurance marketization. Based on big data thinking, this paper studies how to include driving behavior risk into vehicle insurance pricing through the relevant knowledge of the major.

First of all, based on the analysis of traditional vehicle insurance, this paper proposes a new way of rate determination based on driving behavior, so as to optimize the two rate determination modes of "slave vehicle" and "slave person". The traditional car insurance rates are determined from two aspects: one is the "from car" mode, which considers the risk factors including vehicle type, vehicle use and vehicle age; the other is the "from person" mode, which considers the risk factors including driver's age, gender and occupation. Through the analysis of the causes of vehicle insurance claims, it is found that the risk factors considered in these two models are static factors, while the probability of actual traffic accidents is closely related to the dynamic factor of driving behavior. Therefore, the dynamic factor of driving behavior should be fully considered in the determination of vehicle insurance rate.

Secondly, the risk assessment model of vehicle safety related factors is built based on the concept of big data. According to the selection principle of risk factors, the main risk factors affecting driving safety are obtained. According to the index system of AHP, the risk assessment system is established. Then, the entropy weight analytic hierarchy process is used to calculate the index weight of each risk factor, determine the risk assessment score of each risk factor, and establish the risk assessment model.

Finally, the driving behavior risk assessment score is included in the risk factors of vehicle insurance pricing, and the driving behavior coefficient is proposed on the original rate adjustment coefficient, so that the driving behavior can be reflected in the vehicle insurance rate and the vehicle insurance premium. The software and case analysis are used to verify the vehicle insurance pricing mechanism based on large data, and the method of determining the vehicle insurance rate based on driving behavior is proved Scientific.

The research on the risk assessment model of driving behavior based on big data and the determination method of car insurance rate described in this paper can not only promote the healthy development of car insurance industry, but also have theoretical guidance and practical application value for the scientificity, rationality and fairness of car insurance rate, and also have correction and reminder function for driver's driving behavior and promote people's civilized driving.

Key words: big data; automobile insurance; automobile insurance rate

目录

摘要 III

Abstract IV

第1章 绪论 1

1.1 研究背景及意义 1

1.1.1研究背景 1

1.1.2研究目的及意义 2

1.2 国内外研究现状 2

1.2.1 汽车保险市场化研究现状 2

1.2.2 车险费率确定模式和方法现状研究 3

1.2.3 基于大数据的车险费率研究现状 3

1.3 研究内容与方法 4

1.3.1 研究内容 4

1.3.2 研究方法 5

1.4 技术路线图 6

第2章 汽车保险和大数据理论 8

2.1汽车保险 8

2.2 车辆保险的定价 8

2.2.1 车辆保险的理赔 8

2.2.2 车险费率的含义 10

2.2.3 车险费率厘定的流程 10

2.3费率厘定的常用方式 11

2.4 大数据 11

第3章 大数据下的车辆风险因素识别 13

3.1 基于行驶环境的车辆风险因素研究 13

3.1.1 数据来源 13

3.1.2 数据处理 13

3.1.3 研究结果与讨论 13

3.2 基于驾驶行为的车辆风险因素研究 14

3.2.1 新型传感器在车联网中的应用 14

3.2.2 研究结果与讨论 14

3.3 基于行驶里程的车辆风险因素研究 15

第4章 大数据下的车险费率研究 16

4.1车辆事故风险评分指标的构建 16

4.1.1 评分指标构建原则 16

4.1.2 分析交通事故的主要因素 16

4.1.3 建立大数据下的车辆事故风险评分指标体系 17

4.2 事故风险评分指标权重 19

4.2.1 客观赋权法的分析 19

4.2.2 主观赋权法的分析 20

4.3 构建大数据下的车辆事故风险模型 20

4.4 基于大数据评分模型对费率的调整 22

第5章 总结与展望 24

5.1总结 24

5.2展望 24

致谢 26

参考文献 27

第1章 绪论

1.1 研究背景及意义

1.1.1研究背景

我国汽车产业在改革开放和全球化中发展很快，汽车保有量居世界前列。通过查看公安部交通管理局发布的数据，我们可以得知：截至2019年6月，全国汽车保有量达2.5亿辆。其中有1.98亿辆私家车，344万辆新能源汽车。随着生活水平的提高，家家户户都买起了私家车。光是2019年上半年，就有1408万人新领驾驶证。据统计，汽车保有量超过100万辆的城市全国有66个，首都北京超过300万辆。随着买车用车的人越来越多，汽车保险逐渐成为财产保险中的主要产品，机动车保险的经营销售情况也成为整个财产保险的晴雨表。

政府对汽车保险也越来越重视，中国保监会作为相关机构，在2015年印发《关于深化商业车险条款费率管理制度改革的意见》（保监发[2015]18号）。随后，政府部门不断加大简政放权的力度，重视事后和事中的管理，运用现代化与科学化的手段来监管财产保险，得到各试点地区车险消费者普遍认可，市场运行稳中向好。该《意见》放权给市场，保险公司有了一定的自主权，可以用科学的测算并且结合实际市场合理制定调整系数和调整标准，进而确定相应的商业车保险费率。

此举可以使财产保险公司在风险定价方面能够更好地承担责任，有效地整合财产保险公司的风险、经营成本和商业车险费率，鼓励其用服务和价格优势等方面取得经营竞争优势，用专业化、科学化以及精细化的手段来厘定财产保险行业中的商业车险费率。