摘要
在邻域风险最小化原则中提出一种新的定义邻域半径的方法 ,即对任意训练样本点 ,首先利用最佳距离度量近邻法定义一个距离度量 ,并根据这个距离度量来寻找该样本点的最近邻 ,然后依据它们之间的距离来定义邻域半径 ,最后在原有邻域风险最小化算法的基础上建立基于最佳距离度量近邻法的邻域风险最小化算法 .实例表明新提出的算法是正确的。
A new method of defining vicinal radius in vicinal risk minimization (VRM) is proposed.For any training vector,according to optimal distance measure method,a distance measure is proposed.Depending on the distance measure,the vector's nearest neighbour can be found.Then according to the distance between the vector and the vector's nearest neighbour,a new definition of vicinal radius is proposed.At last,a new VRM algorithm based on previous VRM algorithm is set,and two numerical experiments are given to demonstrate the proposed approach's correctness and effectiveness.
出处
《云南大学学报(自然科学版)》
CAS
CSCD
2004年第5期373-377,共5页
Journal of Yunnan University(Natural Sciences Edition)
基金
云南省教育厅基金资助项目 ( 0 2ZY0 11)
云南大学理 (工 )科校级科研项目资助 ( 2 0 0 2Q0 19SL) .
关键词
最佳距离度量近邻法
邻域风险
最小化
支持向量机
vicinal risk minimization
support vector machine
optimal distance measure neighbour