Learning Scoring Functions with Order-Preserving Losses and Standardized Supervision

David Buffoni 1 Clément Calauzènes 1 Patrick Gallinari 1 Nicolas Usunier 1
1 MALIRE - Machine Learning and Information Retrieval
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : We address the problem of designing surrogate losses for learning scoring functions in the context of label ranking. We extend to ranking problems a notion of order preserving losses previously introduced for multiclass classi?cation, and show that these losses lead to consistent formulations with respect to a family of ranking evaluation metrics. An order-preserving loss can be tailored for a given evaluation metric by appropriately setting some weights depending on this metric and the observed supervision. These weights, called the standard form of the supervision, do not always exist, but we show that previous consistency results for ranking were proved in special cases where they do. We then evaluate a new pairwise loss consistent with the (Normalized) Discounted Cumulative Gain on benchmark datasets.
Document type :
Conference papers
Complete list of metadatas

Cited literature [18 references]  Display  Hide  Download
https://hal.archives-ouvertes.fr/hal-00834045
Contributor : Clément Calauzènes <>
Submitted on : Friday, June 14, 2013 - 9:03:01 AM
Last modification on : Thursday, March 21, 2019 - 2:42:18 PM
Long-term archiving on : Sunday, September 15, 2013 - 4:10:09 AM

File

447_icmlpaper.pdf
Publisher files allowed on an open archive

Identifiers

  • HAL Id : hal-00834045, version 1

Collections

UPMC | LIP6 | SORBONNE-UNIVERSITE | SU-SCIENCES

Citation

David Buffoni, Clément Calauzènes, Patrick Gallinari, Nicolas Usunier. Learning Scoring Functions with Order-Preserving Losses and Standardized Supervision. The 28th International Conference on Machine Learning (ICML 2011), Jun 2011, Bellevue, WA, United States. pp.825-832. ⟨hal-00834045⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

219

Files downloads

168

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD