An improved decision tree algorithm based on the attribute set dependency

The decision tree algorithm is the more popular areas of research in data mining and ID3 algorithm is the core algorithm of decision tree algorithm, through research and analysis of the ID3 algorithm, for its shortcoming of multi-value bias interrelated, difficult to remove noise and attribute is not close enough, this study presents attributes set dependence based on rough set theory, doing the attribute reduction considering properties interdependent, thereby removing redundant attributes and the algorithm of attribute set dependence is also given ,at the same time comparing complexity of the algorithm before and after improvement. The draw improved the algorithm is better than before. © 2013 Asian Network for Scientific Information.