Improving the Learning Accuracy of Fuzzy Decision Trees by Direct Back Propagation

Although fuzzy decision trees (FDT) has been a very powerful methodology to extract human interpretable classification rules, it is often criticized to result in poor learning accuracy. In this paper, we propose a methodology to apply back propagation algorithm directly on the fuzzy decision tree structure for improving its learning accuracy without compromising the interpretability. By keeping the tree structure intact, this methodology efficiently tunes the tree parameters with significant increase in the learning accuracy