摘要
从降低时间和空间复杂度的角度出发,针对支持向量机的增量学习问题展开了研究,描述并比较了目前研究与应用较多的几种支持向量机增量学习算法,提出了一种基于壳向量的支持向量机渐进式增量学习算法,仿真实验结果表明:该算法在保证良好的分类精度的前提下,提高了学习效率.
This paper researches incremental learning algorithm of SVM,analyzes the shortcomings of SVM from the time and space complexity.Several incremental learning algorithms of SVM are described and compared,a gradual incremental learning algorithm of SVM based on hull vector is proposed.Simulations show that the algorithm is effective on the premise of ensuring good classification accuracy.
出处
《中南民族大学学报(自然科学版)》
CAS
2011年第3期94-97,共4页
Journal of South-Central University for Nationalities:Natural Science Edition
基金
国家自然科学基金资助项目(60803095)
关键词
增量学习
支持向量机
分类
incremental learning
support vector machine
classification