一种约减型支持向量域数据描述算法.pdfVIP

th Proceedings of the 25 Chinese Control Conference 7-11 August, 2006, Harbin, Heilongjiang 一种约减型支持向量域数据描述算法1 1 2 1,2 赵英刚 ，刘仰光 ，何钦铭 1. 浙江大学 计算机科学与技术学院, 杭州 310027 E-mail: ygz129@163.com 2. 浙江大学宁波理工学院 信息科学与工程分院, 宁波 315100 E-mail: lyg@ 摘 要 支持向量域数据描述(SVDD)是一种单值分类算法，用于将 目标样本与其它非目标样本区分开来，引入 数学中曲率的概念,文中根据分类边界线附近支持向量曲率的大小来对训练集进行约减，提出了一种约减型的 支持向量域数据描述快速训练算法 RSVDD，该算法与传统 SVDD 相比减少了训练时所需的支持向量数目，因而 训练时间大大减少，同时分类性能几乎不受大的影响，该算法在大规模训练样本学习中具有现实的意义。 关键词 支持向量域数据描述，曲率，约减学习 One Kind of Reduced Learning Algorithm for Support Vector Domain Description 1 2 1,2 Yinggang Zhao Yangguang Liu , Qinming He 1. College of Computer Science, Zhejiang University, Hangzhou 310027, China 2. School of Information Science Engineering, Ningbo Institute of Technology, Zhejiang University, Ningbo 315100, China Abstract: As a kind of one-class classification algorithm ，Support Vector Data Description (SVDD) was used to distinguish target objects from outliers objects. By introducing the mathematics concept of curvature, we reduced the training samples according to the value of support vectors locating on the classification boundary, then a reduced learning Support Vector Data Description (RSVDD) algorithm based on the reduced support machines was presented in this paper. Compared with the traditional SVDD, RSVDD o