SVMDT分类器及其在文本分类中的应用研究

SVMDT Classifier and Its Application Research in Text Classification

基于SVM(Support Vectort Machine)理论的分类器已经发展为一种通用的二值分类器,但它不适用于多值的场合。在分析经典的SVM分类算法和决策树分类算法的基础上,提出了将SVM和二叉决策树结合的方法来实现多值分类器(SVMDT),并将其应用于文本分类,实验表明在分类精度和速度上具有良好的性能。

Classifier based on SVM(Support Vectort Machine) theory has been developed as a general-purpose two-class classifier, but it is not suitable for multi-class classification. In the paper two basic classification algorithms are analyzed and combined them into one multi-class classifier(SVMDT).It is applied in text classification and the experimental results show that it has good performance in classification precision and test speed.