交互支持向量机学习算法能解决一些监督学习问

题中学习样本较少的问题, 它以支持向量机(SVM ) 方法为

基础, 将设计分类器变成一个交互的过程, 即: 根据对已知

样本进行的SVM 分类器设计, 主动采样选择“有用”的新样

本, 并进行下一步SVM 分类器的设计。与普通SVM 法相

比, 该方法所需的样本量大大降低, 而且可能达到更好的推

广能力。文本信息过滤问题的实例说明了该算法的有效性。-Interactive support vector machine learning algorithm for supervised learning problems could solve some small problems in the study sample, it is to support vector machine (SVM) method based classifier will be designed into an interactive process, namely: 根据 for known samples for the SVM classifier design, active sampling select " useful"　in the new sample, and the next step SVM classifier design. Compared with the general SVM, the method of sample required is greatly reduced, and may achieve better generalization ability. Text information filtering problem given to illustrate the effectiveness of the algorithm.