大样本的线性可分支持向量机算法-中国科技论文在线.PDF

中国科技论文在线 大样本的线性可分支持向量机算法 # 1 1 2** 乔于 ，易正俊 ，易校石 （1. 重庆大学数学与统计学院，重庆 401331 ； 5 2. 重庆师范大学数学科学学院，重庆 401331 ） 摘要：支持向量机 （Support Vector Machine，SVM）适合于解决高维小样本的分类问题。但 在大数据背景下，支持向量机对于海量样本的学习算法收敛速度较慢，难以在实际中被采用。 论文为了将支持向量机用于海量样本的学习，提出了将大样本随机分成若干小样本的训练数 据集，第一组训练集训练出的支持向量，加入到第二组训练数据，训练出的支持向量加入第 10 三组训练数据，依次类推，直到最后一组训练数据训练的支持向量机和支持向量便是大样本 训练的支持向量机和支持向量。实验结果表明：该支持向量机的算法与解凸二次规划所得到 的支持向量完全一致。 关键词：数理统计学；分类；支持向量机；大样本 中图分类号：O212.4 15 A linear separable support vector machine for large samples 1 1 2 Qiao Yu , Yi Zhengjun , Yi Xiaoshi (1. Chongqing University, School of Mathematics and Statistics, Chongqing 401331; 2. Chongqing normal University, School of Mathematical Science, Chongqing 401331) 20 Abstract : Support vector machine is suitable for solving the classification problem of high dimensional and small samples. However, in the context of big data, support vector machine is slow to converge to the learning algorithm for massive samples, which is difficult to be used in practice. In order to make the SVM that is to be used to study mass samples, this thesis puts forward that large samples are randomly divided into training data set with a number of small samples. The support vector built in the first set of 25 training data set joins the second one, after which the built support vector joins the third one, and so on and so forth, until the SVM and support vector built in the final training data set are the ones