Research Article
An Adaptive Heterogeneous Online Learning Ensemble Classifier for Nonstationary Environments
Table 1
Predictive accuracies (%) of HDES-AD, DDD, OAUE, and AFWE.
| Dataset | HDES-AD | DDD | OAUE | AFWE | Acc | CPU | Acc | CPU | Acc | CPU | Accuracy | CPU |
| Random | 89.12 (1) | 101.2 | 81.27 (2) | 106.2 | 78.69 (4) | 104.6 | 80.33 (3) | 103.4 | SEA | 83.23 (1) | 89.3 | 74.35 (3) | 103.4 | 80.43 (2) | 93.5 | 73.89 (4) | 106.7 | LED | 81.34 (2) | 134.6 | 76.54 (4) | 159.6 | 82.34 (1) | 148.3 | 76.67 (3) | 138.4 | Waveform | 84.47 (1) | 119.3 | 72.87 (3) | 126.2 | 73.27 (3) | 134.2 | 79.65 (2) | 118.2 | Covertype | 85.59 (1) | 108.3 | 78.43 (4) | 114.6 | 82.35 (2) | 113.4 | 71.23 (4) | 139.3 | SpamAssassin | 78.93 (1) | 120.2 | 73.17 (3) | 128.3 | 69.74 (4) | 142.7 | 74.43 (2) | 134.4 | KDD99 | 80.37 (1) | 112.4 | 73.89 (4) | 125.3 | 74.38 (3) | 122.3 | 76.34 (2) | 138.2 | Poker Hand | 76.87 (2) | 104.6 | 78.67 (1) | 115.2 | 7325 (3) | 116.9 | 71.26 (4) | 139.4 | Average ranks | 1.25 | | 3.0 | | 2.75 | | 3.0 | |
|
|