基于知识粗糙度的多变量决策树的构建

A knowledge roughness based approach to multivariate decision tree construction

提出了一种基于粗糙集中知识粗糙度的构建多变量决策树的算法.基本思想是,在选择了核属性作为决策树的根节点后,将选择知识粗糙度最小的那个条件属性作为下一个分枝结点的检验属性.

We propose a new approach to multivariate decision tree construction based on knowledge roughness in rough set instead of information entropy as usual. The basic idea is that, after selecting core attributes as the root of the decision tree, a branching node will then be generated if the corresponding condition attribute provides the smallest knowledge roughness.