Research on Machine Learning Feature Algorithm

The main idea of feature selection is to select feature subsets by removing features that contain little or no relevant information. Feature selection methods can be divided into three categories, filter, encapsulate and embed. Given the large number of feature selection algorithms currently available, in order to be able to properly decide which algorithm to use in a particular situation, it is necessary to propose standards that can be relied upon or determined. This paper is to review some basic feature selection algorithms, compare and classify feature selection methods and algorithms based on the existing theories and experimental results in the literature, and then propose a standard that can be dependent or determined.