Decision tree pruning using backpropagation neural networks

Neural networks have been widely applied to various tasks, such as handwritten character recognition, autonomous robot driving, determining the consensus base in DNA sequences. We describe the use of backpropagation neural networks for pruning decision trees. Decision tree pruning is indispensable for making the overfitting trees more accurate in classifying unseen data. In decision trees, the overfitting can occur when the size of the tree is too large compared to the number of training data. Many methods for decision tree pruning have been proposed, and all of them remove some nodes from the tree to reduce its size. However, some removed nodes may have a significance level or some contribution in classifying new data. Therefore, instead of absolutely removing nodes, our proposed method employs a backpropagation neural network to give weights to nodes according to their significance. Experimental results on twenty domains demonstrate that our method outperforms error-based pruning.