Reliable Multilabel Classification: Prediction with Partial Abstention

Abstract

In contrast to conventional (single-label) classification, the setting of multilabel classification (MLC) allows an instance to belong to several classes simultaneously. Thus, instead of selecting a single class label, predictions take the form of a subset of all labels. In this paper, we study an extension of the setting of MLC, in which the learner is allowed to partially abstain from a prediction, that is, to deliver predictions on some but not necessarily all class labels. We propose a formalization of MLC with abstention in terms of a generalized loss minimization problem and present first results for the case of the Hamming loss, rank loss, and F-measure, both theoretical and experimental.

Cite

Text

Nguyen and Hüllermeier. "Reliable Multilabel Classification: Prediction with Partial Abstention." AAAI Conference on Artificial Intelligence, 2020. doi:10.1609/AAAI.V34I04.5972

Markdown

[Nguyen and Hüllermeier. "Reliable Multilabel Classification: Prediction with Partial Abstention." AAAI Conference on Artificial Intelligence, 2020.](https://mlanthology.org/aaai/2020/nguyen2020aaai-reliable/) doi:10.1609/AAAI.V34I04.5972

BibTeX

@inproceedings{nguyen2020aaai-reliable,
  title     = {{Reliable Multilabel Classification: Prediction with Partial Abstention}},
  author    = {Nguyen, Vu-Linh and Hüllermeier, Eyke},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2020},
  pages     = {5264-5271},
  doi       = {10.1609/AAAI.V34I04.5972},
  url       = {https://mlanthology.org/aaai/2020/nguyen2020aaai-reliable/}
}