摘要
利用粗糙集中决策表的分明矩阵选择多变量决策树的根属性,把信息熵研究属性约简过程中的理论用于节点属性检验和选择,实现多变量决策树的建立。通过实例验证多变量决策树诊断模型较之单变量决策树诊断模型减少了故障信息的冗余性,诊断效率高,结果易于理解。
According to decision table＇s discernibility matrix in rough set theory,root attributes of multivariate decision tree can be selected.Using the theory of attribute reduction based on information entropy to select the node attribute and construct multivariate decision tree.The example shows that the multivariate decision tree model reduces the redundancy of fault diagnosis information,with high efficiency,and easy to be understood than univariate decision tree.
出处
《电脑与信息技术》
2007年第6期21-23,53,共4页
Computer and Information Technology
基金
湖南省教育厅基金资助项目(No.06C125)
关键词
粗糙集理论
分明矩阵
属性约简
条件熵
多变量决策树
单变量决策树
rough set theory
discernibility matrix
attribute reduction
condition entropy
multivariate decision tree
univariate decision tre
