WCCI Performance Prediction Challenge

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

Final #1

Submitted by Gavin Cawley

model selection based on leave-one-out cross-validation, estimation also based on leave-one-out BER (and hence likely to be quite significantly biased).

ADA #045
GINA #123
HIVA #042
NOVA #051
SYLVA #088

Dataset Balanced Error Test guess Guess error Test score Area Under Curve
Train Valid Test Train Valid Test
ada 0.149 0.1542 0.1845 0.1683 0.0162 0.2007 0.9248 0.9259 0.8965
gina 0 0 0.0461 0.0434 0.0027 0.0485 1 1 0.99
hiva 0.018 0.0216 0.2804 0.2475 0.0329 0.3131 0.9979 0.9963 0.7464
nova 0 0 0.0445 0.0436 0.0009 0.0448 1 1 0.9914
sylva 0.0028 0.0029 0.0067 0.0048 0.0018 0.0085 0.9999 0.9998 0.999
Overall 0.034 0.0357 0.1124 0.1015 0.0109 0.1231 0.9845 0.9844 0.9246

This entry is a complete valid challenge entry.