A Unified Loss Function in Bayesian Framework for Support Vector Regression
Abstract
In this paper, we propose a unified non-quadratic loss function for regression known as soft insensitive loss function (SILF). SILF is a flexible model and possesses most of the desirable characteristics of popular non-quadratic loss functions, such as Laplacian, Huber’s and Vapnik’s ε-insensitive loss function. We describe the properties of SILF and illustrate our assumption on the underlying noise model in detail. Moreover, the introduction of SILF in regression makes it possible to apply Bayesian techniques on Support Vector methods. Experimental results on simulated and real-world datasets indicate the feasibility of the approach. 1.
Cite
Text
Chu et al. "A Unified Loss Function in Bayesian Framework for Support Vector Regression." International Conference on Machine Learning, 2001.Markdown
[Chu et al. "A Unified Loss Function in Bayesian Framework for Support Vector Regression." International Conference on Machine Learning, 2001.](https://mlanthology.org/icml/2001/chu2001icml-unified/)BibTeX
@inproceedings{chu2001icml-unified,
title = {{A Unified Loss Function in Bayesian Framework for Support Vector Regression}},
author = {Chu, Wei and Keerthi, S. Sathiya and Ong, Chong Jin},
booktitle = {International Conference on Machine Learning},
year = {2001},
pages = {51-58},
url = {https://mlanthology.org/icml/2001/chu2001icml-unified/}
}