Research Article
Recurrent Adaptive Classifier Ensemble for Handling Recurring Concept Drifts
Table 8
CPU processing time of the algorithms using Hidden Markov Models.
| | Dataset | RACE | CALMID | DUE | SOFEnsemble | AWDOB |
| | Hyperplane | 16.47 (1) | 23.45 (2) | 57.23 (5) | 29.31 (3) | 35.46 (4) | | Stagger | 17.38 (2) | 33.45 (5) | 22.26 (4) | 20.13 (3) | 9.16 (1) | | LED | 22.98 (5) | 15.68 (3) | 13.68 (2) | 12.52 (1) | 17.98 (4) | | SEA | 13.87 (1) | 24.05 (3) | 20.03 (2) | 26.43 (4) | 38.45 (5) | | Random Tree | 30.26 (1) | 38.18 (4) | 37.42 (3) | 31.43 (2) | 47.53 (5) | | Airlines | 63.24 (2) | 60.48 (1) | 82.47 (4) | 80.25 (3) | 84.26 (5) | | KDD99 | 16.68 (1) | 28.14 (4) | 30.53 (5) | 20.23 (2) | 26.68 (3) | | Covertype | 14.64 (3) | 13.48 (2) | 26.48 (4) | 10.41 (1) | 36.17 (5) | | Poker Hand | 20.25 (2) | 28.35 (4) | 16.67 (1) | 26.69 (3) | 30.23 (5) | | Sensor Data | 72.36 (1) | 96.45 (5) | 90.12 (4) | 80.12 (2) | 86.57 (3) | | Average ranks | 1.9 | 3.3 | 3.4 | 2.3 | 4.0 |
|
|
