CC<i>k</i>EL: Compensation-based correlated <i>k</i>-labelsets for classifying imbalanced multi-label data

Abstract

<abstract> <p>Imbalanced data distribution and label correlation are two intrinsic characteristics of multi-label data. This occurs because in this type of data, instances associated with certain labels may be sparse, and some labels may be associated with others, posing a challenge for traditional machine learning techniques. To simultaneously adapt imbalanced data distribution and label correlation, this study proposed a novel algorithm called compensation-based correlated <italic>k</italic>-labelsets (CC<italic>k</italic>EL). First, for each label, the CC<italic>k</italic>EL selects the <italic>k</italic>-1 strongest correlated labels in the label space to constitute multiple correlated <italic>k</italic>-labelsets; this improves its efficiency in comparison with the random <italic>k</italic>-labelsets (RA<italic>k</italic>EL) algorithm. Then, the CC<italic>k</italic>EL transforms each <italic>k</italic>-labelset into a multiclass issue. Finally, it uses a fast decision output compensation strategy to address class imbalance in the decoded multi-label decision space. We compared the performance of the proposed CC<italic>k</italic>EL algorithm with that of multiple popular multi-label imbalance learning algorithms on 10 benchmark multi-label datasets, and the results show its effectiveness and superiority.</p> </abstract>