Research Article
Particle Swarm Optimization Based Selective Ensemble of Online Sequential Extreme Learning Machine
Table 3
Comparison of algorithms for regression problems with sigmoid hidden nodes.
| Datasets | Algorithm | Number of nodes | Number of networks | Training time (s) | RMSE or Accuracy | Testing dev. | Training RMSE | Testing RMSE |
| Abalone | OS-ELM | 25 | | 0.1191 | 0.0758 | 0.0782 | 0.0049 | EOS-ELM | 25 | 5 | 0.5942 | 0.0754 | 0.0775 | 0.0023 | SEOS-ELM (GASEN) | 25 | 5 | 7.3864 | 0.0744 | 0.0760 | 0.0016 | SEOS-ELM (PSOSEN) | 25 | 5 | 4.1528 | 0.0742 | 0.0758 | 0.0015 |
| Mackey-Glass | OS-ELM | 120 | | 0.9827 | 0.0177 | 0.0185 | 0.0018 | EOS-ELM | 120 | 5 | 4.8062 | 0.0176 | 0.0183 | 0.0007 | SEOS-ELM (GASEN) | 120 | 5 | 37.4371 | 0.0172 | 0.0180 | 0.0005 | SEOS-ELM (PSOSEN) | 120 | 5 | 25.1608 | 0.0173 | 0.0179 | 0.0006 |
| California Housing | OS-ELM | 50 | | 0.6871 | 0.1276 | 0.1335 | 0.0035 | EOS-ELM | 50 | 5 | 3.2356 | 0.1280 | 0.1337 | 0.0019 | SEOS-ELM (GASEN) | 50 | 5 | 20.7635 | 0.1242 | 0.1321 | 0.0014 | SEOS-ELM (PSOSEN) | 50 | 5 | 15.6326 | 0.1238 | 0.1323 | 0.0014 |
|
|