【作者】 李欣；

【导师】 李昆仑；

【作者基本信息】 河北大学 ， 电子与通信， 2013， 硕士

【摘要】 近年来，我国保险业务快速增长，服务领域不断拓宽，已成为市场经济风险管理的重要手段。保险业已成为金融和社会保障体系的重要组成部分。然而，保险欺诈也在不断滋生，急剧恶化，造成的社会负面影响，令人堪忧。现已成为影响我国保险业进一步发展的重要瓶颈。本文通过分析现有寿险行业中防保险欺诈的方法，针对现有方法的局限性及当前存在的问题，基于数据挖掘技术，提出一套寿险行业的防欺诈系统方案。本文主要内容包括：1.利用数据挖掘技术，分析了寿险行业防欺诈的现状和国内外现有防欺诈系统的局限性；2.基于太平洋寿险保定中心支公司所提供的数据，设计了寿险防欺诈系统的整体实现方案，包括系统业务模型设计，系统逻辑模型设计和系统物理模型设计；3.根据上述设计方案，研究了分布式环境下的数据预处理过程。提出异构数据的集成策略和脏数据的清洗策略；结合各业务模型特点，在防欺诈系统中成功的应用了数据挖掘中的SPRINT算法。更多还原

【Abstract】 In recent years,the rapid growth of insurance business in China,constantly expandservice areas, has become an important means of risk management of the market economy.The insurance industry has become an important part of the financial and social securitysystem. However, insurance fraud has been breeding, rapid deterioration, caused by thenegative impact of society, worrying. It has become an important bottleneck for the furtherdevelopment of China’s insurance industry.By analyzing the method of anti insurance fraud existing in life insuranceindustry.Aiming at the limitation of existing methods and the existing problems, According tothe data mining technology, the anti fraud system of a life insurance scheme. The maincontent of this paper includes:1. The use of data mining technology, analysis of the life insurance industry antilimitations of present situation and the domestic and foreign existing anti fraud system fraud;2. According to the Pacific Life insurance Baoding branch provides data, The design ofthe whole life insurance anti fraud system, including the design of model systems business,design the system logic model design and physical model3. According to the above design scheme, Study on the process of data preprocessing indistributed environment. The cleaning strategy of heterogeneous data integration strategy andthe dirty data; Combined with the characteristics of the business model, in the anti-fraudsystems successfully applied SPRINT data mining algorithms.更多还原