摘要
决策树是对未知数据进行分类预测的一种方法。自顶向下的决策树生成算法关键是对结点属性值的选择。近似精度是RS中描述信息系统模糊程度的参量,能够准确地刻画粗集。文中在典型的ID3算法的基础上提出了基于RS的算法。该算法基于近似精度大的属性选择根结点,分支由分类产生。该算法计算简单,且分类使决策树和粗集更易理解。
The decision tree is a kind of method to classify to predict for the unknown data. The key of policy- making tree production algorithm which is from top to bottom is the pitch point attribute value. The approximation quality describes parameter of the information system fuzzy degree in RS,and it portrays RS accurately. The algorithm on the basis of RS is proposed in this paper on the basis of typical ID3 algorithm, which chooses root on the basis of approximation quality. Classification produces branch. This algorithm is simple in calculation,and classification makes decision tree and RS easy to be understood.
出处
《计算机技术与发展》
2006年第8期83-84,87,共3页
Computer Technology and Development
基金
安徽省教育厅资助项目(2003kj117)
高校青年基金资助项目(2003)
关键词
粗集
决策树
近似精度
rough set
decision tree
approximation quality
