基于SVM学习方法的分析

Analysis of Studying Method Based on SVM

介绍了增量学习算法、序列最小优化算法、加权支持向量机算法等几种应用于大型数据库,在加快训练速度、降低分类错误率等方面有改进的SVM流行算法.在分析各种算法优缺点的基础上,提出了在线性样本训练、超大规模样本下满足KKT条件的算法是SVM算法的发展方向的观点.

Several kinds of improved support vector machine（SVM） algorithm such as increment learning algorithm, SMO, weighted support vector machine algorithm applied to large scale databases are introduced, to speed up the rate of exercise and to lower the radio of classification mistakes etc are analyzed. Based on the advantages and disadvantages of algorithm, it is pointed out that the development orientation of SVM is to satisfy the algorithm with KKT and with online sample exercise and super large scale databases.