摘要
针对半监督学习中渐进直推支持向量机(Progressive Transductive Support Vector Machines, PTSVM)算法存在训练速度慢,回溯式学习多,学习性能不稳定的问题,提出一种改进的渐进直推支持向量机算法—IPTSVM.该算法利用支持向量的信息选择新标注的无标签的样本点,结合增量支持向量机的迭代更新算法,继承渐进直推支持向量机渐进赋值和动态调整的规则,与PTSVM相比,不仅在一般情况下提高了分类的精度,而且大大提高了算法的速度.在人工模拟数据和真实数据上的实验结果表明了该算法的有效性.
In semi-supervised learning, Progressive Transductive Support Vector Machines （PTSVM） has some drawbacks such as slower training speed, more back learning steps, and unstable learning performance, we propose an improved Progressive Transductive Support Vector Machines learning algorithm -- IPTSVM. It uses the information of support vectors to select new unlabeled samples and combines with the iterative update algorithm of Incremental Support Vector Machines. The algorithm inherits progressive labeling and dynamic adjusting of PTSVM. Compared with PTSVM, the method improves accuracy in general and also enhances speed greatly. Experiment results on synthetic and real data sets show the validity of this algorithm.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2009年第5期142-148,共7页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(60674108,60705004)
河南省科技厅科技计划项目(0821002210091)
关键词
半监督学习
支持向量机
直推式学习
增量学习
semi-supervised learning
support vector machines
transductive learning
incremental learning
