Table 1.
The data sets used (ordered by # of training instances × # of attributes)
| Name | # of training instances | # of attributes | # of test instances | # of classes |
|---|---|---|---|---|
| Mammographic Mass | 673 | 6 | 288 | 2 |
| Car | 1209 | 6 | 519 | 4 |
| Yeast | 1038 | 8 | 446 | 10 |
| German credit | 700 | 20 | 300 | 2 |
| Diabetic retinopathy debrecen | 806 | 20 | 345 | 2 |
| Parkinson speech | 728 | 26 | 312 | 2 |
| Abalone | 2923 | 8 | 1254 | 28 |
| Cardiotocography | 1488 | 23 | 638 | 3 |
| Wine quality | 3425 | 11 | 1469 | 11 |
| KR-vs-KP | 2237 | 37 | 959 | 2 |
| Arrhythmia | 316 | 279 | 136 | 16 |
| Waveform | 3500 | 40 | 1500 | 3 |
| Semeion | 1115 | 256 | 478 | 10 |
| Shuttle | 43,500 | 9 | 14,500 | 7 |
| Secom | 1096 | 591 | 471 | 2 |
| Madelon | 1820 | 500 | 780 | 2 |
| Arcene | 100 | 10,000 | 100 | 2 |
| Convex | 8000 | 784 | 50,000 | 2 |
| KDD09-appentency | 35,000 | 230 | 15,000 | 2 |
| Dexter | 420 | 20,000 | 180 | 2 |
| MNIST basic | 12,000 | 784 | 50,000 | 10 |
| ROT. MNIST + BI | 12,000 | 784 | 50,000 | 10 |
| Amazon | 1050 | 10,000 | 450 | 49 |
| Gisette | 4900 | 5000 | 2100 | 2 |
| CIFAR-10-small | 10,000 | 3072 | 10,000 | 10 |
| Dorothea | 805 | 100,000 | 345 | 2 |
| CIFAR-10 | 50,000 | 3072 | 10,000 | 10 |
Small data sets are shown in the first 16 rows. Large data sets are shown in the last 11 rows