A Survey on Feature Selection: Methods and Challenges

Feature selection is one of the most common preprocessing steps in data mining process and knowledge Engineering. For this purpose, At first the dimension reduction techniques consisting of feature extraction and feature selection are described. Then, Concepts, principles and existing feature selection methods are also described. In this way, procedures of finding selected subsets, including Search-based and non-search based strategies, evaluation criteria will be reviewed; as well as we review supervised approaches related to Feature selection metods. Finally, Some of the problems and challenges facing the current and future feature selection processing are also discussed.

کلیدواژه ها:

Dimension reduction ، feature selection ، Search strategies ، non-search strategies ، generator function ، evaluation criteria