摘要
决策树是数据挖掘中常用的分类技术,其生成的规则便于决策者理解和应用。然而面对较多的属性且含有冗余和噪声属性的记录集生成的决策树时,无法删除冗余属性,造成运算过程复杂。本文旨在通过应用粗糙集理论,将其与决策树方法进行结合,对属性进行约简,降低运算复杂度,并生成相对简化的规则形式,并将其应用到银行个人贷款客户信用评估之中。
Decision tree is one of technologies that are often used in classification. It is easy for decision--makers to understand and apply the rules which are constructed by the decision tree. However, when facing to the decision tree which is of many attributes and constructed by redundant and noisy record sets, it is very difficult to delete redundant attributes by algorithm. This paper introduces the rough sets theory, which combines with the decision tree method, in order to reduce the redundant attribute. All that could reduce the complication of operations and construct relatively simply rules, which can be applied to estimate the credit of bank individual loan customer.
出处
《科技和产业》
2008年第1期57-60,共4页
Science Technology and Industry
基金
黑龙江省博士后基金资助项目(LBH-Z05129)
关键词
数据挖掘
决策树
熵
多变量决策树
粗糙集
data mining
decision tree
entropy
multivariate decision tree
rough sets
