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.

PSMS-SUBSAMPLE-ENSEMBLE

Submitted by H. Jair Escalante

PSMS with subsampling, a ensemble is built with the final SWAR.

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1861 0.2093 0.1806 0.9061 0.8862 0.903 agnostic
gina 0.0038 0 0.0471 1 1 0.99 agnostic
hiva 0.0097 0.0041 0.3013 0.9982 0.9983 0.7729 agnostic
nova 0.0004 0.044 0.0511 1 0.997 0.9896 agnostic
sylva 0.0193 0.009 0.017 0.9969 0.9979 0.9972 agnostic
Overall 0.0439 0.0533 0.1194 0.9802 0.9759 0.9305 agnostic

This entry is a complete agnostic learning entry.