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.

AdaBoost

Submitted by H. Jair Escalante

Boosting ensemble of NNs.

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1404 0.2368 0.1988 0.9449 0.8465 0.883 agnostic
gina 0 0.0285 0.053 1 0.9953 0.9878 agnostic
hiva 0.1153 0.2537 0.2937 0.959 0.8145 0.7689 agnostic
nova 0.0004 0.048 0.0504 1 0.9957 0.9895 agnostic
sylva 0.0072 0.0053 0.0117 0.999 0.9991 0.9983 agnostic
Overall 0.0527 0.1145 0.1215 0.9806 0.9302 0.9255 agnostic

This entry is a Model Selection Game entry.