Feature selection techniques in the context of big data: taxonomy and analysis

Recent advancements in Information Technology (IT) have engendered the rapid production of big data, as enormous volumes of data with high dimensional features grow exponentially in different fields. Therefore, dealing with high-dimensional data creates new challenges in terms of data processing efficiency and effectiveness. To address such challenges, Feature Selection (FS) is among the most utilized dimensionality reduction methods, which is helpful in reducing the high dimensionality of large-scale data by picking up a small subset of related and significant features and eliminating unrelated and redundant features in order to construct effective prediction models. This article provides a comprehensive review of the latest FS approaches in the context of big data along with a structured taxonomy, which categorizes the existing methods based on their nature, search strategy, evaluation process, and feature structure. Moreover, it presents a qualitative analysis of FS methods based on their objective, structure, search strategy, schema, learning task, strengths, and weaknesses. Further, a quantitative analysis is also performed to illustrate the number of publications related to FS based on the timeline, main category, and other sub-categories. An experimental study is also conducted comparing ten methods from different categories using twelve benchmark datasets from the University of California, Irvine (UCI) Machine Learning Repository and Arizona State University (ASU) Feature Selection Repository to evaluate their performance in terms of (accuracy, precision, recall, F-measures, and the number of selected features). Finally, we highlight the research issues and open challenges related to FS to assist researchers in identifying future research directions.