A decision tree based on hierarchical decomposition

Nowadays, people often use decision tree reasoning technique to mining knowledge. The traditional decision tree, which is represented by ID3 proposed by Quinlan (1986), can solve the classification problems very well. But when the number of class increases, the single decision tree produced by ID3 turns out to be complicated and less general. This paper uses the hierarchical decomposition method to deal with multi-class problems by producing the multi-level decision trees. Compared with the single decision tree, the decision tree based on hierarchical decomposition has more advantages in handling the multi-class problems.