Combining Classifier Based on Decision Tree
Yu Yao, Zhongliang Fu, Xianghui Zhao, Cheng Wen-fang · 2009
A new classifier ensemble learning algorithm based on decision tree is proposed. Ensemble learning algorithm is one of the algorithms which have best classification results in many classification algorithms. A decision tree algorithm is a kind of greedy algorithm, it use top-down recursive way to determine the tree structure. The proposed algorithm improved the accuracy of classification by combining the advantage of Boosting algorithm with decision tree. The main idea is to make full use of the advantages of ensemble learning algorithm and decision tree. We introduce the algorithm procession in detail. The proposed algorithm proved that the property which has the smallest classification error rate as of decision tree is equivalent to the branching method of traditional decision tree. The algorithm uses the rapid classification capabilities of decision tree. In the meantime, we take into account the classification accuracy of joint classification. Finally, Experiments with UCI machine learning data sets show the effectiveness of the proposed algorithm.