Universum Learning for Multiclass SVM
Sauptik Dhar, Naveen Ramakrishnan, Vladimir Cherkassky, Mohak Shah · arXiv (Cornell University) · 2016
We introduce Universum learning for multiclass problems and propose a novel formulation for multiclass universum SVM (MU-SVM). We also propose a span bound for MU-SVM that can be used for model selection thereby avoiding resampling. Empirical results demonstrate the effectiveness of MU-SVM and the proposed bound.