A Novel Method for Sea Surface Temperature Prediction Based on Deep Learning

Table 2

The results of p1 on the East China Sea dataset which predict 1 day’s SST value (H = 7).

Methods	RMSE	RMSPE	MAPE	MAE (°C)	ACC
SVR	1.0802	6.2985	3.2420	0.4300	0.9568
SVM	1.7932	13.7467	5.6707	0.6530	0.9380
ARIMA	1.6182	3.2015	5.0987	0.6035	0.9395
BPNN	3.7294	15.2178	14.0391	3.8745	0.8491
RBFNN	2.8274	15.3271	13.8721	2.6102	0.8691
RNN	2.8104	14.4212	12.2157	2.1739	0.8718
GRU	1.5281	7.7845	7.3024	1.5981	0.9325
LSTM	1.7102	8.2459	7.7929	1.7291	0.9283
Updated-LSTM	0.5625	2.3526	1.8956	0.5895	0.9795
GRU-SVM	1.6711	8.3571	7.5042	1.7328	0.9359
WNN	0.6812	2.5032	1.9817	0.6647	0.9632
CEEMDAN-LSTM	1.5397	7.2351	6.9083	1.5402	0.9368
DGCnetwork	0.4471	2.0932	1.5018	0.3218	0.9881