Research Article
A High Accurate Multiple Classifier System for Entity Resolution Using Resampling and Ensemble Selection
Table 1
Accuracy comparison.
| | Dataset | Bagging | DREP | Gentle AdaBoost | REA | RES |
| Synthetic data | breast cancer | 0.8996 ± 0.0002 | 0.9088 ± 0.0002● | 0.9215 ± 0.0003● | 0.9045 ± 0.0001● | 0.9327 ± 0.0002● | innosphere | 0.9647 ± 0.0001 | 0.9732 ± 0● | 0.9946 ± 0.0001● | 0.9832 ± 0○ | 0.9911 ± 0● | dermatology | 0.9541 ± 0.0003 | 0.9676 ± 0.0002● | 0.9649 ± 0.0002● | 0.9588 ± 0.0001 | 0.9825 ± 0.0001● | ILPD | 0.9619 ± 0.0001 | 0.9721 ± 0● | 0.9964 ± 0● | 0.9579 ± 0.0001○ | 0.9929 ± 0.0002● | seismic | 0.9779 ± 0 | 0.9756 ± 0○ | 0.9984 ± 0● | 0.9878 ± 0● | 0.9942 ± 0● | abalone | 0.9812 ± 0 | 0.9812 ± 0.0001 | 0.9880 ± 0● | 0.9833 ± 0 | 0.9940 ± 0● | vote | 0.7592 ± 0.0001 | 0.7495 ± 0.0005 | 0.9515 ± 0.0001● | 0.8155 ± 0.0004● | 0.8447 ± 0.0003● | biodeg | 0.8036 ± 0.0001 | 0.7643 ± 0.0002○ | 0.9643 ± 0.0001● | 0.8155 ± 0.0004● | 0.8214 ± 0.0002● | glass | 0.9255 ± 0.0002 | 0.9275 ± 0.0004● | 0.9804 ± 0.0001● | 0.9216 ± 0.0006○ | 0.9412 ± 0.0003● | diabets | 0.9429 ± 0.0002 | 0.9457 ± 0.0008 | 0.9837 ± 0● | 0.8891 ± 0.0005○ | 0.9457 ± 0 |
| Real data | abt_buy | 0.9417 ± 0 | 0.9444 ± 0● | 0.9372 ± 0.0001○ | 0.9465 ± 0● | 0.9535 ± 0● | amazon_gp | 0.9669 ± 0 | 0.9640 ± 0○ | 0.9525 ± 0.0001○ | 0.9729 ± 0● | 0.9796 ± 0● | dblp_acm | 0.9938 ± 0 | 0.9975 ± 0● | 0.9994 ± 0● | 0.9966 ± 0● | 0.9966 ± 0● | dblp_scholar | 0.9809 ± 0 | 0.9750 ± 0○ | 0.9783 ± 0○ | 0.9817 ± 0 | 0.9848 ± 0● | cora | 0.9407 ± 0.0001 | 0.9650 ± 0● | 0.9550 ± 0● | 0.9465 ± 0● | 0.9600 ± 0● |
| | win/tie/loss | | 9/2/4 | 12/0/3 | 8/3/4 | 14/1/0 |
|
|