Table 4.
Feature selection (FS) results of the final combinations
Data set | Auto-WEKA | Our method | ||
---|---|---|---|---|
# of times FS was conducted | # of times FS selected all features | # of times FS selected no feature | # of times FS was conducted | |
Mammographic Mass | 4 | 4 | 0 | 0 |
Car | 4 | 4 | 0 | 0 |
Yeast | 2 | 2 | 0 | 0 |
German credit | 3 | 3 | 0 | 0 |
Diabetic retinopathy debrecen | 3 | 3 | 0 | 0 |
Parkinson speech | 4 | 3 | 0 | 1 |
Abalone | 4 | 4 | 0 | 0 |
Cardiotocography | 4 | 3 | 0 | 0 |
Wine quality | 4 | 4 | 0 | 0 |
KR-vs-KP | 2 | 2 | 0 | 0 |
Arrhythmia | 3 | 1 | 0 | 0 |
Waveform | 4 | 4 | 0 | 0 |
Semeion | 5 | 5 | 0 | 0 |
Shuttle | 3 | 3 | 0 | 0 |
Secom | 4 | 1 | 0 | 0 |
Madelon | 4 | 1 | 0 | 4 |
Arcene | 3 | 3 | 0 | 0 |
Convex | 4 | 4 | 0 | 0 |
KDD09-Appentency | 5 | 0 | 3 | 0 |
Dexter | 4 | 4 | 0 | 0 |
MNIST basic | 3 | 2 | 0 | 0 |
ROT. MNIST + BI | 2 | 2 | 0 | 0 |
Amazon | 4 | 4 | 0 | 1 |
Gisette | 4 | 4 | 0 | 0 |
CIFAR-10-small | 3 | 3 | 0 | 0 |
Dorothea | 4 | 3 | 1 | 1 |
CIFAR-10 | 3 | 3 | 0 | 0 |
Total | 96 | 79 | 4 | 7 |