Optimal Single-Class Classification Strategies
Abstract
We consider single-class classification (SCC) as a two-person game between the learner and an adversary. In this game the target distribution is completely known to the learner and the learner's goal is to construct a classifier capable of guaranteeing a given tolerance for the false-positive error while minimizing the false negative error. We identify both "hard" and "soft" optimal classification strategies for different types of games and demonstrate that soft classification can provide a significant advantage. Our optimal strategies and bounds provide worst-case lower bounds for standard, finite-sample SCC and also motivate new approaches to solving SCC.
Cite
Text
El-Yaniv and Nisenson. "Optimal Single-Class Classification Strategies." Neural Information Processing Systems, 2006.Markdown
[El-Yaniv and Nisenson. "Optimal Single-Class Classification Strategies." Neural Information Processing Systems, 2006.](https://mlanthology.org/neurips/2006/elyaniv2006neurips-optimal/)BibTeX
@inproceedings{elyaniv2006neurips-optimal,
title = {{Optimal Single-Class Classification Strategies}},
author = {El-Yaniv, Ran and Nisenson, Mordechai},
booktitle = {Neural Information Processing Systems},
year = {2006},
pages = {377-384},
url = {https://mlanthology.org/neurips/2006/elyaniv2006neurips-optimal/}
}