Research Article
Recurrent Adaptive Classifier Ensemble for Handling Recurring Concept Drifts
Table 7
Prediction accuracy of the algorithms using Hidden Markov Models.
| | Dataset | RACE | CALMID | DUE | SOFEnsemble | AWDOB |
| | Hyperplane | 76.47 (2) | 72.53 (3) | 63.24 (5) | 79.63 (1) | 66.38 (4) | | Stagger | 72.32 (1) | 66.58 (4) | 70.33 (2) | 68.28 (3) | 62.48 (5) | | LED | 69.56 (3) | 72.37 (2) | 66.49 (4) | 62.37 (5) | 73.54 (1) | | SEA | 73.48 (1) | 64.47 (5) | 72.43 (3) | 70.27 (2) | 66.45 (4) | | Random Tree | 69.49 (1) | 64.38 (3) | 63.29 (4) | 67.53 (2) | 60.36 (5) | | Airlines | 79.43 (2) | 73.29 (4) | 76.35 (3) | 81.25 (1) | 70.32 (5) | | KDD99 | 72.49 (3) | 76.38 (1) | 74.19 (2) | 69.57 (5) | 70.26 (4) | | Covertype | 81.24 (1) | 75.24 (3) | 72.49 (4) | 78.34 (2) | 69.57 (5) | | Poker Hand | 68.38 (2) | 62.54 (5) | 71.37 (1) | 66.43 (3) | 64.56 (4) | | Sensor Data | 73.48 (1) | 69.42 (3) | 66.38 (4) | 70.43 (2) | 63.27 (5) | | Average ranks | 1.7 | 3.3 | 3.2 | 2.6 | 4.2 |
|
|
