Broad Echo State Network with Reservoir Pruning for Nonstationary Time Series Prediction

Table 5

Evaluation indexes of each model in the US power load dataset.

Model	Training time	SMAPE	MAE	RMSE	R^2
ESN	3.8098	0.0529	1453.4936	1915.5536	0.3696
BLS	0.1380	0.0332	909.3070	1311.3357	0.7045
GRU	169.8198	0.0390	1127.0565	1432.7877	0.6473
LSTM	152.8951	0.0410	1190.1022	1510.7332	0.6079
BESN	77.2083	0.0358	1055.7738	1386.5070	0.6697
BPESN	109.9256	0.0297	875.2953	1175.4368	0.7626