WCCI Performance Prediction Challenge

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

scaling + CA-PCA + ELN

Submitted by Myoung Soo Park

Model is selected by the methods as follows.
1) Scaling up/down each input variable according to the distribution of values.
2) CA-PCA(Class Augmented PCA) is a supervised version of PCA.
3) ELN(Evolutionary Logic Networks) is a classifier which can learn data sets very fast.

Dataset Balanced Error Test guess Guess error Test score Area Under Curve
Train Valid Test Train Valid Test
ada 0.1311 0.1736 0.2308 0.1354 0.0954 0.3262 0.8739 0.806 0.7692
gina 0.0753 0.1178 0.1688 0.0796 0.0892 0.2581 0.9308 0.8952 0.8312
hiva 0.0967 0.3054 0.4352 0.1176 0.3176 0.7527 0.9055 0.7075 0.5648
nova 0.0531 0.052 0.1141 0.053 0.0611 0.1751 0.9437 0.9458 0.8859
sylva 0.0189 0.0202 0.0466 0.019 0.0276 0.0743 0.9806 0.978 0.9534
Overall 0.075 0.1338 0.1991 0.0809 0.1182 0.3173 0.9269 0.8665 0.8009

This entry is a complete valid challenge entry.