决策树分类方法在保险业务中的应用研究.pdfVIP

摘 要 保险是一项风险业务 保险公司的一个重要工作就是进行风险评估 风险 评估对保险公司的正常运作至关重要 保费和保单的设计都需要进行比较详细 的风险分析 在利用数据挖掘工具进行风险分析时 可以利用决策树的方法 在保险公司建立的保单及索赔信息数据库的基础上寻找保单中风险较大的领 域 从而指导保险公司的工作 本文利用 SGI MineSet 中的决策树分类器对医 疗保险数据进行数据挖掘 得出了一些实用的控制风险的规则 并讨论了数据 挖掘方法在保险业务应用中的前景和实效 决策树算法是一种归纳分类算法 它能很好地与超大型数据库结合 并能 处理相关的多种数据类型 连续 离散 布尔 而且其结果容易被人理解 其 分类模式容易转化成分类规则, 由决策树算法形成的分类模型的精确度也较 高 论文在分析了决策树分类的各种典型算法后 重点研究了决策树算法在应 用于大型数据库时的可伸缩性问题 给出了可伸缩性决策树算法结构及流程 关键词 数据挖掘 分类 决策树算法 可伸缩性决策树算法 分类器 保 险 Abstract Insurance is risk business,Insurance company important work is risk estimatation, It is essential to Insurance company,the design of insurance Premium and policy need the detailed Analysis.while making using of data mining tool to Analyse risk,it can adopt decision tree,on the base of policy and claim database,find the relatively risky field., Thus it can guide the work of the insurance company. This paper utilizes SGI MineSet’s Decision tree classfier to mine in Medical Insurance data, And have drawn some rules of practical control risk,at the same time it discuss data mining method’s the prospect and actual effect in the insurance business . The tree algorithm of decision is one kind to sum up the categorised algorithm. it can well combine with the ultra large-scale database, And it can deal with the relevant many kinds of data type(continious, Disperse, Cloth). And its result is easy to be understood, Its categorised mode is easy to transform into a categorised rule, The accuracy of the categorised model formed by the tree algorithm of decision is relatively high too. After analysing various kinds of typical decision tree and classifying algorithms , Key research is applying t