Progressive label enhancement

Zhiqiang Kou, Jing Wang, Yuheng Jia, Xin Geng

Abstract: Highlights•Introduced Progressive Label Enhancement for integrating label distribution and reduction.•Optimized feature space by filtering noise and leveraging manifold structures.•Combined dependency-maximization reduction with label enhancement for better learning.•Validated PLE’s effectiveness on 15 datasets, outperforming six baseline techniques.•Provided a detailed review of LE methods, highlighting PLE’s necessity and contribution.