论文部分内容阅读
大多数决策树构造方法在每个节点上只检验单个属性,这种单变量决策树忽视了信息系统中广泛存在的属性间的关联作用,而且修剪时往往代价很大。针对以上两点,提出了一种基于主成分分析的多变量决策树构造方法,提取信息系统中的若干主成分来构造决策树。实验结果表明,这是一种操作简单,效率很高的决策树生成方法。
Most decision tree construction methods examine only a single attribute at each node. Such a univariate decision tree ignores the interplay of widely existing attributes in an information system and is often costly to prune. Aiming at the above two points, this paper proposes a construction method of multivariate decision tree based on principal component analysis, and extracts some principal components of information system to construct decision tree. The experimental results show that this is a simple and efficient decision tree generation method.