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.

Ensemble of Trees - validation data included

Submitted by Joerg Wichard

Ensemble of Trees
The first time that I included the validation labels

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1731 0.1851 0.183 0.9182 0.8932 0.9029 agnostic
gina 0.0013 0.0536 0.1004 1 0.9915 0.9606 agnostic
hiva 0.124 0.1674 0.3147 0.9474 0.9218 0.7557 agnostic
nova 0.1171 0.134 0.147 0.9546 0.9315 0.9282 agnostic
sylva 0.0041 0 0.0508 0.9997 1 0.9969 agnostic
Overall 0.0839 0.108 0.1592 0.964 0.9476 0.9089 agnostic

This entry is a complete agnostic learning entry.