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 CART models

Submitted by Joerg Wichard

crossvalidated CART models, balanced data sets

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1679 0.2111 0.179 0.9186 0.8763 0.9061 agnostic
gina 0.0124 0.0667 0.0729 0.9996 0.987 0.9791 agnostic
hiva 0.2385 0.3992 0.3165 0.8701 0.7212 0.7439 agnostic
nova 0.1265 0.144 0.1478 0.9622 0.9408 0.9381 agnostic
sylva 0.0045 0.0111 0.007 0.9996 0.9989 0.9988 agnostic
Overall 0.11 0.1664 0.1446 0.95 0.9048 0.9132 agnostic

This entry is a complete agnostic learning entry.