【摘要】 压缩特征向量的维数,在较低维特征空间中进行分类器设计是特征选择与提取的目的。本文介绍了相关概念,给出了常见的几种类别可分性判据,讨论了特征选择的几种常用算法。更多还原

【Abstract】 To compress the dimensions of characteristic vector and make the design of classificatory in the characteristic space with lower dimensions is the purpose of characteristic selection and extraction. In this paper, we introduce some conceptions correlative to it, give some kinds of dissociable basis and we put the stress on discussing some algorithms most in use of the characteristic selection.更多还原