摘要
当样本数量大到计算机内存中放不下时,常规支持向量机方法就失去了学习能力,为了解决这一问题,提高支持向量机的训练速度,文章分析了支持向量机分类的本质特征,根据支持向量机分类仅与支持向量有关的特点,提出了一种适合于支持向量机增量学习的快速循环算法(PFI SVM),提高了支持向量机的训练速度和大样本学习的能力,而支持向量机的分类能力不受任何影响,取得了较好的效果。
The general training method of SVM will not work when the amount of training samples is too large to be put into the RAM of computer. In order to solve this problem and improve the speed of training SVM,the nature characteristics of SVM is analyzed in this paper. A fast iteration algorithm suitable for incremental learning for SVM is presented. This algorithm improves the speed of training support vector machine and the ability of learning greatly in the situation of large amount of samples while the classifying ability of SVM is unaffected. Several simulations demonstrate the effectiveness of this approach.
出处
《计算机应用》
CSCD
北大核心
2003年第10期12-14,17,共4页
journal of Computer Applications
基金
国家自然科学基金资助项目 (6 0 2 750 2 0 )
关键词
数据挖掘
支持向量机
分类
循环算法
data mining
support vector machine(SVM)
classification
iteration algorithm
