中文摘要：

SVM决策树是解决多分类问题的有效方法之一,由于分类器组合策略不同,构成的决策树构型以及分类精确度也各有差异。提出基于欧氏距离的SVM决策树构造方法,通过两种欧氏距离组合策略,生成不同构型的SVM决策树。实验结果表明,采用组合策略二的SVM决策树分类器相比组合策略一,具有更高的分类精度和更短的训练及测试时间。

英文摘要：

SVM decision tree is a quite effective method to solve classification-related problems.Due to different cluster strategies,there will be different tree configurations,and this effects the accuracy of the classification.In order to solve this problem,puts forward an euclidean distance-based support vector machine decision tree construction method.Through adopting two different strategies of Euclidean distance algorithm,we have constructed two different types of SVM decision tree structure.Experimental result shows that it is more accurate to adopt the SVM decision tree based on strategy II configuration algorithm than the one based on strategy I configuration algorithm.