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

Submitted by Joerg Wichard

ensemble of binary decision trees, no extras, only
balanced trainings data.

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1674 0.2209 0.1955 0.9218 0.8687 0.9013 agnostic
gina 0.0063 0.0919 0.0921 0.9997 0.9762 0.9641 agnostic
hiva 0.1432 0.2934 0.3029 0.9487 0.7876 0.7583 agnostic
nova 0.0281 0.098 0.1041 0.9989 0.9555 0.9483 agnostic
sylva 0.0048 0.0176 0.0104 0.9991 0.9979 0.9979 agnostic
Overall 0.07 0.1444 0.141 0.9737 0.9172 0.914 agnostic

This entry is a complete agnostic learning entry.