Particle Swarm Optimization Based Selective Ensemble of Online Sequential Extreme Learning Machine

Table 6

Comparison of algorithms for classification problems with RBF hidden nodes.

Datasets	Algorithm	Number of nodes	Number of networks	Training time (s)	RMSE or Accuracy		Testing dev.
New-thyroid	OS-ELM	20		0.0118	93.45%	89.92%	0.0702
	EOS-ELM	20	15	0.1682	93.87%	89.86%	0.0428
	SEOS-ELM (GASEN)	20	15	1.9745	94.53%	91.05%	0.0332
	SEOS-ELM (PSOSEN)	20	15	1.2315	94.68%	91.32%	0.0315
Image segmentation	OS-ELM	180		2.6702	94.98%	91.92%	0.0324
	EOS-ELM	180	5	13.2174	94.39%	91.35%	0.0148
	SEOS-ELM (GASEN)	180	5	128.3215	95.62%	95.06%	0.0085
	SEOS-ELM (PSOSEN)	180	5	90.2856	96.02%	95.24%	0.0079
Satellite image	OS-ELM	400		45.2702	93.62%	89.54%	0.0056
	EOS-ELM	400	10	448.1347	93.86%	89.37%	0.0034
	SEOS-ELM (GASEN)	400	10	4263.1406	94.61%	90.38%	0.0022
	SEOS-ELM (PSOSEN)	400	10	3145.8528	94.85%	90.57%	0.0019