This paper extends the idea of Universum learning [18, 19] to regression problems. We propose new Universum-SVM formulation for regression problems that incorporates a priori knowledge in the form of additional data samples. These additional data samples or Universum belong to the same application domain as the training samples, but they follow a different distribution. Several empirical comparisons are presented to illustrate the utility of the proposed approach.
@article{arxiv.1605.08497,
title = {Universum Learning for SVM Regression},
author = {Sauptik Dhar and Vladimir Cherkassky},
journal= {arXiv preprint arXiv:1605.08497},
year = {2016}
}