摘要
对大规模训练样本的支持向量机训练问题进行探索,提出了一种基于正交表的并行学习算法.这种方法通过求解一些相互独立的小的训练问题来求解大的训练问题,采用多处理机可求解大规模的训练问题.
Explores the training problems of support vector machine with large training pattern set, and a new parallel algorithm based on orthogonal array is presented. This method breaks a large training problem into some small independent problems. Then the large problems can be solved via solving those small problems individually. Thus we will be able to use this algorithm and the computer with many CPU to calculate large problems
出处
《云南大学学报（自然科学版）》
CAS
CSCD
北大核心
2006年第2期93-97,共5页
Journal of Yunnan University(Natural Sciences Edition)
基金
国家自然科学基金资助项目(10271103)
云南大学理(工)科校级科研资助项目(2002Q019SL)
关键词
正交表
并行计算
SVM
HBSVM
orthogonal array
parallel compute
support vector machine （SVM）
HBSVM
