摘要
使用支持向量机算法直接求海量数据的模糊分类系统是相当困难的。为了解决这个问题,提出了基于邻域原理设计模糊分类系统的方法。将支持向量机的理论建立在距离空间上,设计出了计算支持向量的邻域算法;利用所求的支持向量,基于平分最近点方法设计出了求分类超平面的算法,求出模糊分类系统,该算法优于基于支持向量机直接求模糊分类系统的方法。实验结果说明,该方法可有效地解决对海量数据的模糊分类系统的设计问题。
It is rather difficult to design a fuzzy massive data classification system by using support vector machine directly algorithm. To solve this problem, a method is proposed to design fuzzy classification system based on neighborhood theory. Vapnik's support vector machine theory is constructed on the distance space, the algorithm is designed to compute support vectors; Make use of the support vector, the algorithm is designed to calculate optimal classification hyper plane based on the method of dividing equally point recent, then, to get fuzzy classification system, that calculate way better than the method of getting directly fuzzy classification system according to support the vector machine; The result of experiment illustrates the problem of designing fuzzy massive data classification system is effectively solved by using this method.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第17期4065-4067,4115,共4页
Computer Engineering and Design
基金
国家973重点基础研究发展计划基金项目(2004CB318103)
湖南省科技计划基金项目(05JT1039)
长沙市科技计划基金项目(K041039-12)。
关键词
邻域原理
支持向量
支持向量机算法
模糊分类系统
海量数据
neighborhood theory
support vector
support vector machine algorithm
fuzzy classification system
massive data