摘要
决策树是数据挖掘中的一种重要分类方法。在此以粗糙集理论中的正域为启发式函数,设计了一种新的、有效的决策树构造方法。该算法具有较大的灵活性,能从测试属性空间逐次删除已使用过的属性。避免对这些属性进行重复测试,减少测试空间,降低了树的复杂性,从而提高了分类效率。最后,实例验证了算法的可行性与有效性。
Decision tree is an important method to solve classification problems in data mining. In this paper, the positive region in the rough set is used as the heuristic function to design a novel and effective method to build decision tree. The algorithm with flexibility can avoid repeatedly testing these attributes by gradually deleting those used attributes, reduce the testing attributes space and the complexity of the tree, thus improve the classification efficiency. Furthermore, an example is given to verify the feasibility and effectiveness of the algorithm.
出处
《河池学院学报》
2008年第5期71-74,共4页
Journal of Hechi University
关键词
决策树
决策表
粗糙集
正域
decision tree
decision table
rough set
positive region