COLT
2008
Springer
14 years 2 months ago
2008
Springer
We study the potential benefits to classification prediction that arise from having access to unlabeled samples. We compare learning in the semi-supervised model to the standard, ...