为了对高维的基因表达数据进行有效的降维,本研究采用基于Fisher准则的思想对特征空间进行特征组合,从而得到新的特征空间。实验结果表明,在无法对数据再进行降维的原特征空间中,经过线性变换的方法得到的新特征空间可继续对其数据进行降维。通过这种方法,能将四川省医学科学院动物研究所所提供的抑郁症大鼠DNA的基因表达数据有效地从4353维降到1393维。%In order to get effective dimension reduction in high-dimensional gene expression data,this paper is to introduce the experiment of re-combining the original feature space to get the new feature dimension by applying Fisher principle.The results show that,through the linear transformation,the new feature dimension can continue to reduce the dimension of its data,which can not be reduced in the original one.In this way,the date of rats' DNA gene expression data provided by Sichuan Academy of Medical Sciences and Institute of Zoology can be reduced from 4353 dimension to 1393 dimension effectively.
展开▼