Constraint Classification for Multiclass Classification and Ranking
Abstract
The constraint classification framework captures many flavors of mul- ticlass classification including winner-take-all multiclass classification, multilabel classification and ranking. We present a meta-algorithm for learning in this framework that learns via a single linear classifier in high dimension. We discuss distribution independent as well as margin-based generalization bounds and present empirical and theoretical evidence showing that constraint classification benefits over existing methods of multiclass classification.
Cite
Text
Har-Peled et al. "Constraint Classification for Multiclass Classification and Ranking." Neural Information Processing Systems, 2002.Markdown
[Har-Peled et al. "Constraint Classification for Multiclass Classification and Ranking." Neural Information Processing Systems, 2002.](https://mlanthology.org/neurips/2002/harpeled2002neurips-constraint/)BibTeX
@inproceedings{harpeled2002neurips-constraint,
title = {{Constraint Classification for Multiclass Classification and Ranking}},
author = {Har-Peled, Sariel and Roth, Dan and Zimak, Dav},
booktitle = {Neural Information Processing Systems},
year = {2002},
pages = {809-816},
url = {https://mlanthology.org/neurips/2002/harpeled2002neurips-constraint/}
}