ADT: A decision tree algorithm based on concepts

Abstract

In this paper, a new method named Alternative Decision Trees (ADT) for the generation of decision trees is introduced. This proposed method generated a decision tree based in concepts of minimum covering, obtained of concepts or properties of each class described in the data set. Starting from the limitations of others decision tree algorithms, we worked in the development of new techniques that allow us to improve those limitations. A simplification of these concepts, and a split criterion are introduced. Besides, the performance of ADT method is presented.