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.

Liknon

Submitted by Erinija

Liknon aka LP-SVM, ref: Bhattacharya C. et all, Signal Processing 83(4): 729-743, 2003

Dataset Balanced Error Area Under Curve  
Train Valid Test Train Valid Test
ada 0.1878 0.1947 0.1865 0.8946 0.8669 0.8949 agnostic
gina 0.0493 0.1267 0.1338 0.978 0.9298 0.9391 agnostic
hiva 0.1274 0.3676 0.3439 0.9285 0.7295 0.7025 agnostic
nova 0.0199 0.108 0.1001 0.9975 0.9645 0.9612 agnostic
sylva 0.0224 0.0291 0.026 0.9948 0.997 0.9947 agnostic
Overall 0.0814 0.1652 0.1581 0.9587 0.8975 0.8985 agnostic

This entry is a complete agnostic learning entry.