Active k-labelsets ensemble for multi-label classification

The random k-labelsets ensemble (RAkEL) is a multi-label learning strategy that integrates many single-label learning models. Each single-label model is constructed using a label powerset (LP) technique based on a randomly generated size-k label subset. Although RAkEL can improve the generalization capability and reduce the complexity of the original LP method, the quality of the randomly generated label subsets could be low. On the one hand, the transformed classes may be difficult to separate in the feature space, negatively affecting the performance; on the other hand, the classes might be highly imbalanced, resulting in difficulties in using the existing single-label algorithms. To solve these problems, we propose an active k-labelsets ensemble (ACkEL) paradigm. Borrowing the idea of active learning, a label-selection criterion is proposed to evaluate the separability and balance level of the classes transformed from a label subset. Subsequently, by randomly selecting the first label or label subset, the remaining ones are iteratively chosen based on the proposed criterion. ACkEL can be realized in both the disjoint and overlapping modes, which adopt pool-based and stream-based frameworks, respectively. Experimental comparisons demonstrate the feasibility and effectiveness of the proposed methods.