基于支持向量机的嵌入拒识代价的手写字符识别研究  

SVM-based Hand-written Character Recognition with Reject Option

作　　者：黄垚[1] 刘思颂[1] 孔瑞[1] 

机构地区：[1]成都理工大学,四川成都610059

出　　处：《电子质量》2011年第4期5-7,共3页Electronics Quality

摘　　要：支持向量机(SVM-Support Vector Machine)是在统计学习理论基础上发展起来的一种新的机器学习方法。传统的SVM是基于两类问题的,而实际需要解决的一般是多类问题。因此,将SVM应用于多类问题对挖掘SVM的应用潜力将具有非常重要的意义。此研究针对SVM在手写字符中的作用所提出的拒识代价和误分类代价是一个新的研究点,且更具有实际的研究意义。Support Vector Machine（SVM-Support Vector Machine） is in statistical learning theory developed on the basis of a new machine learning methods.SVM traditional is based on the two types of problems,and the actual needs of the general solution are a wide range of issues.So the question of SVM applied to many types of mining SVM on the potential applications will be of great significance.The design put forward by the Rejection costs and the price of wrong classification is a new study,and has more research practical significance.

关 键 词：手写字符 支持向量机 拒识代价 误分类代价 概率点 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]