简述特征选择

Brief Description of Feature Selection

特征选择是通过寻找对于目标函数有突出贡献的特征来达到降维的效果,该方法希望可以尽可能多的去掉冗余特征,能够更加准确合理地解释这些数据。研究者们对于特征选择的研究历史也比较悠久,从而特征选择也变得越来越准确、有效。本文介绍特征选择概念之后,简单综述了特征选择在方法和理论上的发展,并且重点介绍了支持向量机在特征选择上的应用。

Feature selection is to reduce dimension by finding features that contribute significantly to the ob-jective function. This method hopes to remove redundant features as much as possible and inter-pret these data more accurately and reasonably. Researchers have a long history of research on feature selection, so feature selection is becoming more and more accurate and effective. After in-troducing the concept of feature selection, this paper briefly reviews the development of feature se-lection methods and theories, and focuses on the application of support vector machines in feature selection.