The Application Of Svm To Decision Tree Induction

Decision tree induction learning is the most important branch of machine learning. Usually, the heuristic algorithms are usually used to construct a decision tree. So researching kinds of heuristic algorithms to induce a decision tree with high accuracy becomes a focus. Support vector machine(SVM) is a kind of learning algorithm for small samples based on statistical learning theory. Considering the relationship between the classification margin of support vector machine and the generalization capability, SVM can be applied into the decision tree induction. The large margin can be used as the heuristic information of decision tree in order to improve its generalization capability. In this paper, the decision tree induction learning using large margin as the heuristic information is researched based on the statistic learning theory and support vector machine.In order to improve the generalization capability of the decision tree, the large margin theory in the SVM is applied into the decision tree induction in this paper by analyzing the support vector machine and its inverse problem deeply. First, the basic problem of support vector machine and its fast solving algorithm are discussed, the inverse problem of support vector machine and its solving is also discussed based on this, the solving algorithm based on k-means clustering is proposed. Second, the application of SVM to decision tree induction is discussed and the algorithm that using large margin as the heuristic information to induce a decision tree is proposed and then the algorithm is analyzed. Finally, the experiment process is introduced, comparing with the binary decision tree based on minimum entropy heuristic information, the results show the algorithm proposed in this paper is valid.