Complexity to Forecast Flood: Problem Definition and Spatiotemporal Attention LSTM Solution

Table 5

Comparisons on Changhua dataset with several comparative methods. Note that bold texts indicate best performance among comparative methods.

Models	RT+3	RT+6	RT+9	RA	MT+3 (%)	MT+6 (%)	MT+9 (%)	MA (%)	DT+3	DT+6	DT+9	DA
FCN	56.87	88.36	129.7	91.63	14.22	22.09	32.41	22.91	0.84	0.64	0.31	0.60
SVM	60.89	86.34	110.6	85.93	15.22	21.59	27.64	21.48	0.82	0.67	0.53	0.67
LSTM	66.11	84.49	109.2	86.60	16.53	21.12	27.30	21.65	0.79	0.68	0.54	0.67
XAJ model	77.30	68.31	69.25	71.62	19.26	17.09	17.45	17.93	0.65	0.70	0.71	0.69
STA-LSTM	52.37	78.36	94.55	75.09	13.09	19.59	23.64	18.77	0.85	0.71	0.62	0.73