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.

mixed tree ensembles

Submitted by Joerg Wichard

balanced data sets
decision trees selected via
modified bagging

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1663 0.2241 0.1905 0.925 0.8724 0.9037 agnostic
gina 0.0041 0.0411 0.061 0.9999 0.9913 0.9834 agnostic
hiva 0.138 0.3514 0.2977 0.949 0.7481 0.7605 agnostic
nova 0.1051 0.128 0.1251 0.9644 0.9488 0.9407 agnostic
sylva 0.0036 0.0111 0.007 0.9997 0.999 0.9986 agnostic
Overall 0.0834 0.1511 0.1362 0.9676 0.9119 0.9174 agnostic

This entry is a complete agnostic learning entry.