A comparison between fuzzy-ID3 and OFFSS-based fuzzy-ID3

In this paper, a method of combining fuzzy decision tree with optimal fuzzy-valued feature subset selection (OFFSS) is proposed. This algorithm can first select a feature subset from the entire feature space, and then constructs the fuzzy decision tree on the selected feature subset. This work conducts some experiments to compare the proposed algorithm with fuzzy-ID3. The experiment results show that fuzzy decision trees on feature subset are superior to that on entire feature space in terms of speed and accuracy for classification.