论文部分内容阅读
针对传统支持向量机中存在原始数据量过大导致训练速度太慢的问题,同时考虑到非支持向量对支持向量机的训练性能无影响,且影响支持向量机性能的支持向量往往位于边界的特点,提出一种提取边界向量的支持向量机算法.数值实验表明:改进算法在保证支持向量机分类能力的前提下,有效提高了支持向量机的分类效率.
Aiming at the problem that the original data is too large, the training speed is too slow. Considering that the non-support vector has no effect on the training performance of the SVM, and the support vector which often affects the performance of the SVM is often located at the boundary , A support vector machine (SVM) algorithm for extracting boundary vectors is proposed.The numerical experiments show that the improved algorithm can effectively improve the classification efficiency of SVM under the premise of guaranteeing the classification ability of SVM.