Scientific Reports

Table 2 Comparative data of evaluation indexes of training set and test set of each model.

From: Decomposition prediction and optimal ensemble strategy improve river dissolved oxygen prediction accuracy

		Train				Test
Stations	Models	MAE	MSE	RMSE	R²	MAE	MSE	RMSE	R²
A	LSTM	0.399	0.285	0.534	0.843	0.306	0.171	0.413	0.735
	SVR	0.307	0.184	0.429	0.898	0.259	0.124	0.353	0.807
	MLP	0.437	0.340	0.583	0.812	0.354	0.214	0.463	0.668
	C-LSTM	0.258	0.114	0.337	0.937	0.264	0.107	0.327	0.834
	C-SVR	0.235	0.100	0.317	0.945	0.205	0.073	0.269	0.887
	C-MLP	0.213	0.085	0.292	0.953	0.192	0.068	0.260	0.895
	Optimal ensemble	0.210	0.084	0.290	0.954	0.186	0.065	0.256	0.898
B	LSTM	0.267	0.186	0.431	0.912	0.232	0.122	0.350	0.929
	SVR	0.220	0.146	0.382	0.931	0.198	0.109	0.331	0.936
	MLP	0.270	0.185	0.430	0.912	0.220	0.115	0.338	0.933
	C-LSTM	0.245	0.097	0.312	0.954	0.209	0.074	0.272	0.957
	C-SVR	0.170	0.065	0.255	0.969	0.163	0.054	0.232	0.969
	C-MLP	0.169	0.063	0.250	0.970	0.164	0.054	0.232	0.969
	Optimal ensemble	0.158	0.059	0.243	0.972	0.138	0.046	0.214	0.973

Back to article page

Search

Advanced search

Quick links