Stable matching-based two-way selection in multi-label active learning with imbalanced data

Multi-label active learning (MLAL) reduces the cost of manual annotation for multi-label problems by selecting high-quality