A Comparative Study of Feature Selection Algorithms in Data Mining

Abstract: : Data mining is the method of extraction of related data from a collection of large dataset. Mining of a fastidious data related to a concept is done on the basis of the feature of the data. The accessing of these features thus for data retrieval can be termed as the feature selection mechanism. Different types of feature selection methods are being used. Feature selection methods in data Mining problem aim at selecting a subset of the features, which illustrate the data in order to acquire a more necessary and compact representation of the available information. The preferred subset has to be small in size and must maintain the information that is most useful for the specific application. This paper try to analyze Feature Selection algorithms clearly with the purpose to examine strengths and weaknesses of some widely used Feature Selection methods.

Keywords: Feature selection algorithm, Euclidian distance, T-test, Information gain, Markov blanket filter.