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.

single model

Submitted by reference (gcc)

LS-SVM, model selection via PRESS statistic, RBF kernel, single model trained on entire training set, but not using the validation labels.

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1576 0.2046 0.1887 0.927 0.8729 0.9 agnostic
gina 0 0.0442 0.0499 1 0.9955 0.9892 agnostic
hiva 0.0388 0.2819 0.2708 0.9983 0.7604 0.7771 agnostic
nova 0.0004 0.048 0.0619 1 0.9942 0.9852 agnostic
sylva 0.002 0.0049 0.0081 1 0.9991 0.9988 agnostic
Overall 0.0398 0.1167 0.1159 0.9851 0.9244 0.9301 agnostic

This entry is a complete agnostic learning entry.