SSLearn is an open-source Python library designed to support semi-supervised learning (SSL), with a focus on wrapper algorithms and restricted set classification (RSC), offering a unique set of tools compatible with Scikit-Learn. It provides researchers and practitioners with a flexible, extensible platform to experiment with, compare, and develop SSL methods, filling a gap in existing libraries by including RSC and enabling easier adoption of SSL techniques in real-world applications.

This comprehensive review analyzes 128 semi-supervised ensemble learning methods published since 2013, categorizing them by approach, ensemble type, and base classifier, and evaluating their experimental protocols, dataset usage, and statistical rigor. The study highlights the dominance of wrapper methods (especially co-training), the underuse of statistical validation, and the need for more reproducible research, while offering future directions for advancing semi-supervised ensemble learning in information fusion.

This study introduces a novel semi-supervised classification method that incorporates automatically generated pairwise constraints—Must Link (ML) and Cannot Link (CL)—from video data to improve animal re-identification accuracy. By applying these constraints frame-by-frame and combining them with standard classifiers, the proposed methods significantly outperform traditional classifiers, constrained clustering, and other semi-supervised learning approaches across five animal video datasets.