数据挖掘论文：基于支持向量机的保险业务数据挖掘系统的设计和实现.docVIP

数据挖掘论文：基于支持向量机的保险业务数据挖掘系统的设计与实现 【中文摘要】作为一家大型的保险公司,中国人寿保险公司河北分公司已经具有非常完善的信息化管理系统,但为了适应日趋激烈的市场竞争环境,适应以产品为中心到以客户为中心的经营理念转变,充分利用历年积累的大量宝贵的信息资源,对客户信息进行挖掘,建立并完善保险公司客户关系,进行客户关系数据挖掘,提高保险公司的核心竞争力势在必行。本文着重就客户关系建立应用模型,分析了支持向量机算法在数据挖掘方面的应用,并对算法作了应用扩展,并将该算法引入客户特征提取分析的应用中来。最后通过建立数据挖掘模型,完成了本系统。在系统构建过程中,主要工作体现在以下几点:1)三层的体系架构,使得系统能够处理多源数据,进行多种类型数据的格式转化,从而适应现在公司中多系统,数据混乱的情形;2)采用支持向量机算法进行数据挖掘,在选取客户多特征进行训练时,通过核函数的选取,能够进行高维空间映射处理;3)建立数据挖掘模型,利用实际数据进行检验,使得模型具有高度的准确性和实用性。论文最后就该系统在中国人寿保险公司河北分公司的试用情况进行了介绍,并就客户忠诚度利用系统挖掘模型进行了分析。 【英文摘要】As a large-scale insurance company, Hebei Branch of China Life Insurance Company already has a very perfect information management system, but to meet the increasingly intensive market competition, it is imperative for the insurance company to enhance its core competitiveness by converting product-orientated managerial ideology to customer-orientated managerial ideology, capitalizing on its immense valuable information resources acquired in years, mining the customer information and customer relationship data in the effort to establish and improve its customer relationships.The thesis emphasizes establishing the application model for customer relationships, analyses and extends the application of support vector machine algorithm in data mining aspect, and introduces the algorithm into the application of customer characteristics analyses. The system is finally finished by establishing data mining models.Works involved in the system construction process falls into the following three categories:1). 3-tier system framework, enable the system to process multi source data and to convert the multi pattern data formats; it thereupon adapts the company to multi-system and multi-data situations.2). Adopt support vector machine algorithm in data mining; manage high dimensional space mapping by choosing kernel function in the customer multi characteristics training.3). Establish the data mining model; test the practical data to get high accuracy and practicability.The thesi