uEFS: An efficient and comprehensive ensemble-based feature selection methodology to select informative features

• DOI: 10.1371/journal.pone.0202705
• Publication Date: 2018-08-28T17:48:58Z
• Authors: Maqbool Ali; Syed Imran Ali; Dohyeong Kim; Taeho Hur; Jaehun Bang; Sungyoung Lee; Byeong Ho Kang; Maqbool Hussain
• Keywords: Feature (linguistics); Benchmark (surveying); Univariate

Feature selection is considered to be one of the most critical methods for choosing appropriate features from a larger set items. This task requires two basic steps: ranking and filtering. Of these, former necessitates all features, while latter involves filtering out irrelevant based on some threshold value. In this regard, several feature with well-documented capabilities limitations have already been proposed. Similarly, also nontrivial, as it designation an optimal cutoff value so properly select important list candidate features. However, availability comprehensive approach, which alleviates existing provides efficient mechanism achieving results, major problem. Keeping in view these facts, we present univariate ensemble-based (uEFS) methodology informative input dataset. For uEFS methodology, first propose unified scoring (UFS) algorithm generate final ranked following evaluation set. defining points remove subsequently (TVS) subset that are deemed classifier construction. The evaluated using standard benchmark datasets. extensive experimental results show our proposed competitive accuracy achieved (1) average around 7% increase f-measure, (2) 5% predictive compared state-of-the-art methods.