Uncertain Decision Tree Inductive Inference

Induction is the process of reasoning in which General rules are formulated based on limited observations of recurring phenomenal patterns. Decision tree learning is one of the most widely used and practical inductive methods, which represents the results in a tree scheme. Various decision tree algorithms have already been proposed, such as CLS, ID3, Assistant and C4.5. These algorithms suffer from some major shortcomings. In this paper, after discussing the main limitations of the existing methods, we introduce a new decision tree induction algorithm, which overcomes all the problems existing in its counterparts. We also illustrate the advantages and the new features of the proposed method. The experimental results will show the effectiveness of the method in comparison with other methods existing in the literature.