Support Vector Machines with Example Dependent Costs
Abstract
Classical learning algorithms from the fields of artificial neural networks and machine learning, typically, do not take any costs into account or allow only costs depending on the classes of the examples that are used for learning. As an extension of class dependent costs, we consider costs that are example, i.e. feature and class dependent. We present a natural cost-sensitive extension of the support vector machine (SVM) and discuss its relation to the Bayes rule. We also derive an approach for including example dependent costs into an arbitrary cost-insensitive learning algorithm by sampling according to modified probability distributions.
Cite
Text
Brefeld et al. "Support Vector Machines with Example Dependent Costs." European Conference on Machine Learning, 2003. doi:10.1007/978-3-540-39857-8_5Markdown
[Brefeld et al. "Support Vector Machines with Example Dependent Costs." European Conference on Machine Learning, 2003.](https://mlanthology.org/ecmlpkdd/2003/brefeld2003ecml-support/) doi:10.1007/978-3-540-39857-8_5BibTeX
@inproceedings{brefeld2003ecml-support,
title = {{Support Vector Machines with Example Dependent Costs}},
author = {Brefeld, Ulf and Geibel, Peter and Wysotzki, Fritz},
booktitle = {European Conference on Machine Learning},
year = {2003},
pages = {23-34},
doi = {10.1007/978-3-540-39857-8_5},
url = {https://mlanthology.org/ecmlpkdd/2003/brefeld2003ecml-support/}
}