Rough set based decision tree
Jinmao Wei, Dao Zheng Huang, Shuqin Wang, Zhuyang Ma · 2003
Decision tree is widely used in machine learning. One important step in construction of a decision tree is how to select appropriate attributes as nodes of the tree. There are many approaches to selection of attributes. In this paper, we present a new approach to selection of attributes for construction of decision tree based on the rough set theory. Decision trees constructed by the presented approach tend to have simpler structure and higher classification accuracy from a statistical point of view than the entropy-based method under some conditions. Some data sets from UCI machine learning database repository are then used to test the two methods, which from application perspective instantiates the performance of rough set-based method. In the paper we also give an algorithm in a recursive form for the construction of decision tree.