An Efficient Feature Selection with an Enhanced Supervised Term-Weighting Scheme in Multi-Class Text Classification

Term weighting scheme and feature selection are two fundamental components in text classification (TC) systems, particularly in high-dimensional, multi-class, and imbalanced settings. Term weighting schemes aim to improve document representation by emphasizing discriminative terms across classes, while feature selection (FS) seeks to reduce dimensionality, eliminate irrelevant and redundant features, and enhance classification efficiency and effectiveness. However, most existing studies focus on FS independently of the term-weighting strategy used during document representation, thereby limiting the potential benefits of their interaction. This study addresses this gap by pursuing two main objectives. First, it employs an enhanced supervised term-weighting scheme, namely MTF-MICF, to construct a more stable and class-discriminative document representation, especially for imbalanced data. Second, it investigates the effectiveness of integrating this scheme with a filter-based FS approach using Information Gain (IG) at various levels of dimensionality reduction to assess the contribution of enhanced term weighting to the FS process. Extensive experiments were conducted across 19 benchmark multi-class text datasets. The performance was evaluated using F1-score and classification accuracy with the three prominent classifiers (MNB, SVM, and LR). The experimental results demonstrate that the proposed approach consistently outperforms conventional methods, achieving significant and stable improvements in both representation quality and classification performance. These findings confirm that enhanced supervised term weighting can serve as an effective supporting mechanism for FS in high-dimensional TC tasks.