Abstract: In this paper,an improved K-means clustering algorithm is applied in the training of Support Vector Machine（SVM）.Based on the clustering algorithm,the steps for incrementally training SVM are given.Moreover,a new criterion of eliminating non-informative samples in the training process is developed.The result of pattern classification experiment shows that the application of clustering algorithm in SVM not only greatly reduces the training time of SVM,but also further improves the classification ability of it.

Key words: K-means clustering, incremental training, Support Vector Machine（SVM）

摘要： 将一种改进的K均值聚类算法应用于支持矢量机（SVM）的训练。基于这一改进的聚类算法,设计了SVM的增量式训练步骤,并给出了在训练过程中删除无用样本的的方法。模式分类的实验结果表明,这种改进的K均值聚类算法在SVM中的应用不仅大幅度地缩短了SVM的训练时间,而且进一步提高了它的分类能力。

关键词: K均值聚类算法, 增量训练, SVM