路径跟踪线性规划向量机  

Path following method on linear programming support vector machine

作　　者：陈晨[1] 陈琴[1] 苏一丹[1] 朱茜[1] CHEN Chen CHEN Qin SU Yi-dan ZHU Qian(School of Computer and Electronics Information, Guangxi University, Nanning 530004, Chin)

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004

出　　处：《计算机工程与设计》2017年第8期2132-2136,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61363027)

摘　　要：研究路径跟踪线性规划支持向量机(path following linear programming support vector machine,PF-LPSVM)分类算法,利用路径跟踪法求解线性规划的高效性,提高线性规划支持向量机在大规模数据集上的学习效率。给出线性规划支持向量机的模型并将其标准化,导出用路径跟踪法求解线性规划向量机的关键公式,给出完整的算法流程。在随机数据集及UCI数据集上,将所提算法与LibSVM和牛顿法线性规划向量机(Newton-LPSVM,N-LPSVM)做比较,实验结果表明,所提算法用路径跟踪法提高LPSVM的学习效率是可行的,其适用于大规模数据集的学习。The path following linear programming support vector machine （PF-LPSVM） classification algorithm was studied. The advantage of efficiency of the path following method in solving linear programming was used to improve the learning efficiency of LPSVM on large scale data sets. The model of linear programming support vector machine was given and standardized. The key formula of linear programming vector machine was derived using path following method, and the complete algorithm flow was given. On the random data set and UCI data set, the proposed algorithm was compared with LibSVM and Newton method linear programming vector machine （Newton-LPSVM, N-LPSVM）. Experimental results show that the algorithm proposed is feasible to improve the learning efficiency of LPSVM using path following method, which is suitable for large-scale data sets.

关 键 词：路径跟踪法 线性规划 支持向量机 分类算法 标准化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]