Research Article
The Optimisation for Local Coupled Extreme Learning Machine Using Differential Evolution
Table 4
Regression results of ELM, OSFuzzyELM, LC-ELM, and ELC-ELM.
| Dataset | Algorithm | Training (%) | Training time | Testing (%) | Testing time | Number of nodes/rules | RMSE | SD | RMSE | SD |
| Abalone | ELM | 0.0761 | 0.0012 | 0.0172 | 0.0776 | 0.0027 | 0.0016 | 25 | OSFuzzyELM | 0.0741 | 0.0015 | 1.3026 | 0.0766 | 0.0023 | 0.0406 | 5 | LC-ELM | 0.0754 | 0.0011 | 0.4524 | 0.0767 | 0.0018 | 0.2699 | 25 | ELC-ELM | 0.0754 | 0.0013 | 537.1114 | 0.0743 | 0.0025 | 0.2278 | 25 |
| Autoprice | ELM | 0.0879 | 0.0101 | 0.0016 | 0.1205 | 0.0190 | 0.0009 | 15 | OSFuzzyELM | 0.0401 | 0.0052 | 0.0624 | 0.1096 | 0.0078 | 0.0252 | 3 | LC-ELM | 0.0757 | 0.0044 | 0.0031 | 0.0842 | 0.0087 | 0.0016 | 15 | ELC-ELM | 0.0805 | 0.0059 | 37.1688 | 0.0769 | 0.0048 | 0.0025 | 15 |
| Bodyfat | ELM | 0.0678 | 0.0038 | 0.0062 | 0.1295 | 0.0191 | 0.0047 | 50 | OSFuzzyELM | 0.0790 | 0.0038 | 0.0686 | 0.1264 | 0.0323 | 0.0031 | 2 | LC-ELM | 0.0661 | 0.0038 | 0.0562 | 0.1243 | 0.0243 | 0.0218 | 50 | ELC-ELM | 0.0680 | 0.0013 | 193.9529 | 0.1129 | 0.0115 | 0.0224 | 50 |
| Computer | ELM | 0.0339 | 0.0008 | 0.3354 | 0.0408 | 0.0044 | 0.0406 | 125 | OSFuzzyELM | 0.0257 | 0.0006 | 83.6555 | 0.0346 | 0.0044 | 0.1950 | 15 | LC-ELM | 0.0345 | 0.0016 | 2.2511 | 0.0407 | 0.0033 | 0.9812 | 125 | ELC-ELM | 0.0279 | 0.0003 | | 0.0288 | 0.0005 | 1.0868 | 125 |
| CPU | ELM | 0.0476 | 0.0066 | 0.0016 | 0.0865 | 0.0584 | 0.0008 | 10 | OSFuzzyELM | 0.0284 | 0.0035 | 0.0499 | 0.0659 | 0.0339 | 0.0031 | 3 | LC-ELM | 0.0416 | 0.0070 | 0.0047 | 0.0582 | 0.0203 | 0.0016 | 10 | ELC-ELM | 0.0394 | 0.0069 | 29.3719 | 0.0360 | 0.0092 | 0.0025 | 10 |
| Housing | ELM | 0.0793 | 0.0047 | 0.0062 | 0.0929 | 0.0093 | 0.0016 | 50 | OSFuzzyELM | 0.0645 | 0.0043 | 0.3011 | 0.0924 | 0.0163 | 0.0094 | 5 | LC-ELM | 0.0711 | 0.0035 | 0.1076 | 0.0884 | 0.0091 | 0.0390 | 50 | ELC-ELM | 0.0682 | 0.0040 | 281.3338 | 0.0807 | 0.0127 | 0.3768 | 50 |
|
|