WCCI Performance Prediction Challenge

The challenge is over, but a new challenge is on-going using the same datasets, check it out!

SVM-LOO

Submitted by Olivier Chapelle

RBF Kernel, model selection by minimizing the leave-one-out error.
Performances estimated with Leave-One-Out.

Dataset Balanced Error Test guess Guess error Test score Area Under Curve
Train Valid Test Train Valid Test
ada 0.1375 0.1639 0.184 0.1674 0.0166 0.2006 0.9368 0.9141 0.9005
gina 0 0.0506 0.068 0.0717 0.0038 0.0715 1 0.9797 0.98
hiva 0.0263 0.3695 0.2918 0.2412 0.0506 0.3424 0.9998 0.6625 0.7502
nova 0.002 0.088 0.0737 0.0757 0.002 0.0749 1 0.9653 0.9746
sylva 0.0042 0.0154 0.0137 0.0098 0.0039 0.0176 0.9998 0.9991 0.9987
Overall 0.034 0.1375 0.1262 0.1132 0.0154 0.1414 0.9873 0.9041 0.9208

This entry is a complete valid challenge entry.