ML Anthology

Direct Optimization of Margins Improves Generalization in Combined Classifiers

Llew Mason, Peter L. Bartlett, Jonathan Baxter
NeurIPS 1998 pp. 288-294
/neurips/1998/mason1998neurips-direct/

Abstract

Cumulative training margin dis(cid:173) tributions for AdaBoost versus our "Direct Optimization Of Margins" (DOOM) algorithm. The dark curve is AdaBoost, the light curve is DOOM. DOOM sacrifices significant training er(cid:173) ror for improved test error (hori(cid:173) zontal marks on margin= 0 line)_

PDF NeurIPS Semantic Scholar

Cite

Text

Mason et al. "Direct Optimization of Margins Improves Generalization in Combined Classifiers." Neural Information Processing Systems, 1998.

Markdown

[Mason et al. "Direct Optimization of Margins Improves Generalization in Combined Classifiers." Neural Information Processing Systems, 1998.](https://mlanthology.org/neurips/1998/mason1998neurips-direct/)

BibTeX

@inproceedings{mason1998neurips-direct,
  title     = {{Direct Optimization of Margins Improves Generalization in Combined Classifiers}},
  author    = {Mason, Llew and Bartlett, Peter L. and Baxter, Jonathan},
  booktitle = {Neural Information Processing Systems},
  year      = {1998},
  pages     = {288-294},
  url       = {https://mlanthology.org/neurips/1998/mason1998neurips-direct/}
}