Variable Selection Methods, Comparison and their Applications in Machine Learning: A Review

In the past decade, the availability of voluminous and high-dimensional datasets has immensely emerged with continuous technological innovations to foster new ways to gather and analyze data. Thus, feature selection has become a challenging task in areas of application like text classification, data analysis, prediction, information retrieval etc. Knowledge extraction using machine learning models usually depends on the quality and quantity of data that they operate on. Feature selection is one of the core concepts to identify and remove irrelevant as well as redundant information that may impact accuracy or have no impact on the results. Feature selection methods are discussed in this review paper along with practical limitations. Subsequently, the workflow to solve a feature selection problem is also elaborated with feature selection methods. In feature selection, many surveys and empirical assessments were performed in many areas like classification, prediction, regression, and clustering, respectively