WCCI Performance Prediction Challenge

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

SNB(CMA) + 100k F(2D) tv

Submitted by Marc Boulle

Same method as SNB(CMA) tv

Test of statistical and computational scalability

100 000 features constructed for each dataset
Each one is the sum of two randomly selected initial features

Dataset Balanced Error Test guess Guess error Test score Area Under Curve
Train Valid Test Train Valid Test
ada 0.145 0.1561 0.1723 0.165 0.0073 0.1793 0.9304 0.9287 0.9149
gina 0.0272 0.016 0.0733 0.077 0.0037 0.0767 0.9963 0.9946 0.9772
hiva 0.2227 0.2873 0.308 0.317 0.009 0.3146 0.8532 0.7981 0.7542
nova 0.0364 0.032 0.0813 0.106 0.0247 0.106 0.9924 0.9914 0.9749
sylva 0.0035 0.0029 0.0057 0.008 0.0023 0.008 0.9996 0.9996 0.9991
Overall 0.087 0.0988 0.1281 0.1346 0.0094 0.1369 0.9544 0.9425 0.9241

This entry is a complete valid challenge entry.