@inproceedings{bd6252d890b74ca08367f450bc79488b,
title = "Conformal Rule-Based Multi-label Classification",
abstract = "We advocate the use of conformal prediction (CP) to enhance rule-based multi-label classification (MLC). In particular, we highlight the mutual benefit of CP and rule learning: Rules have the ability to provide natural (non-)conformity scores, which are required by CP, while CP suggests a way to calibrate the assessment of candidate rules, thereby supporting better predictions and more elaborate decision making. We illustrate the potential usefulness of calibrated conformity scores in a case study on lazy multi-label rule learning.",
author = "Eyke H{\"u}llermeier and Johannes F{\"u}rnkranz and \{Loza Menc{\'i}a\}, Eneldo",
year = "2020",
doi = "10.1007/978-3-030-58285-2\_25",
language = "English",
isbn = "978-3-030-58284-5",
volume = "12325",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "290--296",
editor = "Ute Schmid and Diedrich Wolter and Franziska Kl{\"u}gl",
booktitle = "KI 2020",
}