The Test of Time Award honors the paper from ECML PKDD 2009 with the highest impact in the field. The award is sponsored by European Research Center for Information Systems (ERCIS).
Classifier Chains for Multi-label Classification
Jesse Read, Bernhard Pfahringer, Geoff Holmes, Eibe Frank (The University of Waikato)
The widely known binary relevance method for multi-label classification, which considers each label as an independent binary problem, has been sidelined in the literature due to the perceived inadequacy of its label-independence assumption. Instead, most current methods invest considerable complexity to model interdependencies between labels. This paper shows that binary relevance- based methods have much to offer, especially in terms of scalability to large datasets. We exemplify this with a novel chaining method that can model label correlations while maintaining acceptable computational complexity. Empirical evaluation over a broad range of multi-label datasets with a variety of evaluation metrics demonstrates the competitiveness of our chaining method against related and state-of-the-art methods, both in terms of predictive performance and time complexity.
Praise by the Award Chairs: The well-known transformation of a multi-label problem into a binary problem for each label has the disadvantage that the labels might correlate. Passing label correlation information along a chain of classifiers counteracts this deficiency while maintaining the ease of implementation and understanding. The paper is extremely well recognised by 537 citations of the conference paper and 981 of the subsequent journal version.