中文摘要：

为了完成分类学习，传统的支持向量机基于带标记信息的经验数据归纳出一个通用的决策函数。而转导支持向量机则不同．它考虑包含测试集在内的所有数据信息并致力于最小化测试样本的分类错误数。在已有的2类分类方法的基础上构造了直接求解多类分类问题的的转导支持向量机。

英文摘要：

While regular Support Vector Machines try to induce a general decision function based on the labled empirical datas for a classfication learning task,Transductive Support Vector Machines take into account a particular test set and try to minimize misclassfications of just those paiticular ex- amples. The Transductive Support Vector Machines for multiclass classification learning are constructed directly based on the well studied methods of binary classfication.