PASCAL Agnostic Learning
vs.
Prior Knowledge
IJCNN07

The challenge is now over. But it remains open for post-challenge submissions!


IMPORTANT: Entries made since February 1st 2007 might be using validation data, now available for training.

mixed ensembles (balanced data)

Submitted by Joerg Wichard

using more than 20 CV-Partitions

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1657 0.1997 0.1843 0.9175 0.8763 0.9046 agnostic
gina 0.0016 0.0507 0.0673 1 0.9887 0.9809 agnostic
hiva 0.1971 0.2988 0.3137 0.9145 0.7486 0.7486 agnostic
nova 0.1265 0.134 0.1443 0.9463 0.9158 0.9329 agnostic
sylva 0.0083 0.0185 0.0132 0.9992 0.9982 0.9982 agnostic
Overall 0.0998 0.1403 0.1445 0.9555 0.9055 0.913 agnostic

This entry is a complete agnostic learning entry.