The study on technologies for feature selection
Jun Zhao, Guoyin Wang, Zhongfu Wu, Hong Tang, Hua Li · 2003
Feature selection is crucial to machine learning. The diversity of learning systems makes it impossible to thoroughly solve the problem of feature selection with a specific tool or approach. In this paper, methods, models, and relevant technologies concerning feature selection are inspected thoroughly, several typical algorithms for feature selection are described, and, finally, some guidelines are given to choose an appropriate algorithm to select features according to the characteristics of a learning system.