论文部分内容阅读
传统决策树算法有好的可理解性,而支持向量机集成方法有好的分类性能。本文基于二者的优点提出了一种基于支持向量机集成的决策树分类算法,首先在样本集合上训练支持向量机集成分类器。然后随机生成一定数目的特征向量并加以标注,构造辅助样本。最后在辅助样本的帮助下生成新的决策树。由于最终结果由决策树给出,所以本文算法具有很好的可理解性。此外在辅助样本的帮助下,算法的分类性能也得到提高。在UCI标准数据集与新闻文本数据集上的实验充分验证了本文算法的合理性。
The traditional decision tree algorithm has good comprehensibility, and the SVM integration method has good classification performance. Based on the advantages of the two, this paper proposes a decision tree classification algorithm based on support vector machine (SVM) integration. Firstly, SVM classifier is trained on the sample set. Then a certain number of eigenvectors are randomly generated and labeled to construct the auxiliary samples. Finally, a new decision tree is generated with the help of the auxiliary samples. Since the final result is given by the decision tree, the algorithm in this paper has good comprehensibility. In addition, with the help of auxiliary samples, the classification performance of the algorithm is also improved. Experiments on UCI standard dataset and news text dataset fully verify the rationality of the proposed algorithm.