多重筛选逐步回归特征选择法及其应用

A Multiple Sieving Stepwise Regression Method for Feature Selection and Its Application

  • 摘要: 本文介绍了利用统计模式识别中的多重筛选逐步回归特征选择法,对五类白血球进行特征选择,并对所选用的特征用训练集和测试集进行识别效果检验,得到了训练集和检测集上的两个混淆矩阵.

     

    Abstract: In this paper,a multiple sieving stepwise regression method for feature selection in statstical pattern recognition is introduced.The method is applied to featureselection for five classes of leucocyte,and the recognition effectiveness of the selected features is examined with train set and test set.Two mixed-matrices in thetrain set and test set are obtained.

     

/

返回文章
返回