%Aigaion2 BibTeX export from Knowledge Engineering Publications
%Friday 17 December 2021 11:56:18 PM

@INPROCEEDINGS{ln:MLCaggregation,
     author = {Nguyen, Vu-Linh and H{\"{u}}llermeier, Eyke and Rapp, Michael and Loza Menc{\'{\i}}a, Eneldo and F{\"{u}}rnkranz, Johannes},
     editor = {Appice, Annalisa and Tsoumakas, Grigorios and Manolopoulos, Yannis and Matwin, Stan},
   keywords = {Combine then Predict, Ensembles of Multilabel Classifiers, F-measure, Hamming loss, Predict then Combine, Subset 0/1 loss},
      month = oct,
      title = {On Aggregation in Ensembles of Multilabel Classifiers},
  booktitle = {Discovery Science},
     series = {Lecture Notes in Computer Science},
     volume = {12323},
       year = {2020},
      pages = {533--547},
  publisher = {Springer International Publishing},
    address = {Cham},
       isbn = {978-3-030-61527-7},
        url = {https://arxiv.org/abs/2006.11916},
        doi = {https://doi.org/10.1007/978-3-030-61527-7_35},
   abstract = {While a variety of ensemble methods for multilabel classification have been proposed in the literature, the question of how to aggregate the predictions of the individual members of the ensemble has received little attention so far. In this paper, we introduce a formal framework of ensemble multilabel classification, in which we distinguish two principal approaches: "predict then combine" (PTC), where the ensemble members first make loss minimizing predictions which are subsequently combined, and "combine then predict" (CTP), which first aggregates information such as marginal label probabilities from the individual ensemble members, and then derives a prediction from this aggregation. While both approaches generalize voting techniques commonly used for multilabel ensembles, they allow to explicitly take the target performance measure into account. Therefore, concrete instantiations of CTP and PTC can be tailored to concrete loss functions. Experimentally, we show that standard voting techniques are indeed outperformed by suitable instantiations of CTP and PTC, and provide some evidence that CTP performs well for decomposable loss functions, whereas PTC is the better choice for non-decomposable losses.}
}