首页 | 本学科首页   官方微博 | 高级检索  
     

分离错误最小化的极大熵方法
引用本文:姜翠萍,董玉林,高莘莘. 分离错误最小化的极大熵方法[J]. 辽宁师范大学学报(自然科学版), 2006, 29(2): 160-162
作者姓名:姜翠萍  董玉林  高莘莘
作者单位:中国海洋大学,青岛学院,公共课教学部,山东,青岛,266300;大连理工大学,应用数学系,辽宁,大连,116024
摘    要:分离错误最小化是支持向量机的基本问题之一.一种形式是最小化分离错误点的偏离和,这是一个不可傲优化问题,笔者提出用极大熵函数将其转化成可微凸规划问题来处理,得到原问题的近似最优解。

关 键 词:分离错误最小化  分类超平面  极大熵方法  凸函数
文章编号:1000-1735(2006)02-0160-03
收稿时间:2005-04-20
修稿时间:2005-04-20

A Maximum Entropy Method for Misclassification Minimization
JIANG Cui-ping,DONG Yu-lin,GAO Shen-shen. A Maximum Entropy Method for Misclassification Minimization[J]. Journal of Liaoning Normal University(Natural Science Edition), 2006, 29(2): 160-162
Authors:JIANG Cui-ping  DONG Yu-lin  GAO Shen-shen
Affiliation:1. Basic Courses of Teaching and Research Department, Qingdao College,Ocean University of China, Qingdao 266300,China;2. Department of Applied Mat hematics, Dalian University of Technology, Dalian 116024, China
Abstract:Misclassification Minimization is a fundamental problem of machine learning. It can be stated by a way of minimizing the sum of violations of misclassified points. It is NP-complete. The objective function is not differentiable. In this paper,a convex entropy function is used to solve the nondifferentiable optimal problem,and the approximate solution is achieved by this convex programming.
Keywords:misclassification minimization   separation hyperplane   maximum entropy method   convex function
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号