摘要
在不对判别变量进行处理的条件下,对传统的距离判别方法进行改进,提出一种新的判别方法,试图解决复杂球形数据的判别问题,以提高判别的正确率。通过实例表明,该方法的判别效果良好,能较好地处理复杂球形数据的判别问题。
This paper presents a new discriminant method through improving the traditional distance discriminant method and tries to improve the accuracy of discriminant problems of complex spherical data in condition of not reducing variables. The example shows that this method has good discriminant effect, and can effectively deal with the discriminant problems of complex spherical data.
出处
《江南大学学报(自然科学版)》
CAS
2011年第6期745-748,共4页
Joural of Jiangnan University (Natural Science Edition)
基金
泉州师范学院校级项目(2008KJ015)
关键词
距离判别
FISHER判别
判别准则
马氏距离
distance discriminant analysis, Fisher discriminant analysis, discriminant rule, Mahalanobis distance