LGSep 29, 2016
Universum Learning for Multiclass SVM
arXiv:1609.09162v15 citations
Originality Incremental advance
AI Analysis
This work addresses multiclass classification problems, but appears incremental as it extends existing Universum SVM methods to the multiclass setting.
The authors tackled the problem of multiclass classification by introducing Universum learning to multiclass SVM, proposing a novel formulation and a span bound for model selection, and demonstrated its effectiveness empirically.
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.