《重庆人保财险私家车保险风险管理研究》由会员分享，可在线阅读，更多相关《重庆人保财险私家车保险风险管理研究（54页珍藏版）》请在金锄头文库上搜索。
1、重庆人保财险私家车保险风险管理研究重庆大学专业硕士学位论文学生姓名：岳鸿川指导教师：李 勇专业学位：MBA重庆大学经济与工商管理学院二OO九年四月Research on Insurance Risk Management for private car of Chongqing PICC A Thesis Submitted to Chongqing Universityin Partial Fulfillment of the Requirement for theDegree of Master of ProfessionByYue HongchuanSupervised by Li Yo
2、ngProfession: Master of Business AdministrationCollege of Economics and Business Administration of Chongqing University, Chongqing, ChinaApril 2009 / 文档可自由编辑打印摘 要据统计，截至2008年底全国汽车保有量为6400万余俩，其中私家车4100余万辆，占汽车保有量的64.53%，而1990年中国私家车保有量仅为82万辆，占汽车保有量14.8%。私家车的增长过于迅猛，受各种因素的影响，整个重庆私家车保险市场处于集体亏损的临界状态，重庆人保财险和
3、行业其他竞争主体一样在私家车的承保环节存在很多不足之处，信息化水平较低，数据的积累和管理薄弱、信息共享机制十分欠缺、承保过程中存在众多的违规操作以及核保内容简单，环节单一，难以形成标准化等一系列问题，导致在后期提供服务更多的反而是价值最低的客户，给公司带来最高价值的客户得不到任何服务。这些问题不仅有客观的因素，也有保险公司主观的原因。为了经营好私家车保险，防范未来风险，保险公司必须深度分析私家车的风险状况，细分私家车保险市场，建立更为完善的风险识别和风险管理机制，才能更好更快的发展。本文根据重庆人保财险公司私家车承保风险管理状况，提取了过去具有代表性的客户数据资料样本，并对样本进行了SAS聚类
4、分析和分类模型研究，通过聚类软件工具的作用下发现了高风险、中风险和低风险三类客户分类的存在，同时通过分类模型的处理我们发现了影响客户风险大小的特征因素，以此来鉴别其他客户的风险大小。这样的工作思路和结果显示了重庆人保财险公司在数据挖掘这方面存在的问题和不足，分析了公司的风险管理现状，在基于对私家车保险客户风险类别进行细分的基础上，对不同的客户群体制定相关的风险管理策略，提升公司的经营效益、促使高风险客户做好风险防范措施，增强社会和谐，以使重庆人保财险公司在激烈的车险市场竞争中做好做强，追求卓越，做行业领先者。关键词：数据挖掘，聚类分析，分类预测，私家车保险，风险管理ABSTRACTAccord
5、ing to statistics, amount of car with insurance have been more than two of 64000000 at the end of 2008 cars in the country for, of which private cars have been more than 41000000, accounting for 64.53 percent of all. While private cars in China in 1990 just to maintain the volume of 820000, accounti
6、ng for 14.8 percent of cars volume.The whole private car insurance market in china is critical state of the collective loss for the impact of various risk factors because of the rapidly growth. The part of the underwriting for PICC insurance company of Chongqing are many inadequacies and is a lower
7、level of information and is weak on data accumulation and management and lack of information-sharing mechanisms. And it exist a large number of violations of nuclear security operation simple and single link in underwriting process. It is difficult to form a series of issues of standardization as to
8、 make more service for minimum value of the customer in the provision of services at a later stage, and make less service for the highest value to the companys customers. These problems are not only objective factors but also subjective reasons for insurance companies. In order to run a private car
9、insurance and to guard against future risks, insurance companies must deeply analysis of the risk pro private cars and car breakdown insurance market. Then establish o a more comprehensive risk identification and risk management mechanisms in order to make the company better and faster development.I
10、n this paper ,we extracted some representative customer data sample according to the state of risk management of PICC insurance company of Chongqing. And we research the data sample by the model of cluster and classification, and then we can found high-risk customer, medium risk customer and low-ris
11、k customer. At the same time we can find characteristic factors which influence the degree of risk in order to distinguish the other customers degree of risk .This thought and result display the problem and lack about data mining for PICC insurance company of Chongqing .It analysis company state of
12、risk management and make relation risk management strategy for different customer based on the customer segmentation of private car customer with insurance .And the we can promote operating efficiency and do a good job for high-risk customer in risk prevention measures to enhance social harmony so a
13、s to make PICC insurance company of Chongqing pursuit excellence and be a leader of the industry in the fierce market competition in the fierce auto insurance market competition .Keywords：Data Mining, Cluster Analysis, Classification and Predict , Insurance of Private Car, Risk Management目 录中文摘要I英文摘
14、要II1 绪 论11.1 研究的目的和意义11.2 论文研究的主要内容及框架22 风险管理与数据挖掘的相关理论概述32.1 风险管理理论32.1.1 风险的定义32.1.2 风险识别52.1.3 风险分析52.1.4 风险评估62.1.5 风险管理62.2 数据挖掘理论72.2.1 数据挖掘概述72.2.2 数据挖掘方法82.2.3 数据挖掘在保险业客户风险管理中的应用92.2.4 聚类分析方法（以K均值为例）92.2.5 分类理论（以决策树为例）102.3 SAS Enterprise Miner简介122.4 WEKA平台的介绍143 重庆人保财险公司私家车承保风险管理现状分析163.1
15、重庆人保财险公司简介163.2 重庆人保财险公司私家车承保风险管理现状163.3 重庆人保财险公司私家车客户数据分析与利用的现状与问题183.4 基于数据挖掘的私家车承保风险管理的总体思路193.5 小结194 基于私家车承保风险的分类预测模型分析214.1 问题描述214.2 数据采集与预处理214.2.1 数据模型214.2.2 数据预处理234.3 聚类分析244.3.1 数据模型244.3.2 数据预处理244.3.3 聚类分析254.3.4 聚类模型评估294.4 分类304.4.1 数据模型304.4.2 数据预处理314.4.3 分类模型的建立324.4.4 模型评估374.4.5 模型在客户分类预测中的应用384.5 小结395 聚类分析和分类预测结果在承保风险管理中的应用405.1 应用概述405.2 承保管理策略415.3 理赔管理策略415.4 综合服务策略425.5 其他处理策略435.6 小结436 结 论44致 谢45参考文献461 绪 论1.1 研究的目的和意义社会经济的发展，人民生活水平的提升，私家车保险市场蛋糕越来越
