Methodology for Developing Hydrological Models Based on an Artificial Neural Network to Establish an Early Warning System in Small Catchments

Table 5

Performance statistics of the ANN model during evaluation: mean absolute error (MAE) for the data classes.

Target data versus output data cm			S15	S30	S60
Maximum absolute error cm			98.24	111.34	117.37
Class	Percentage %	Water level class limits; cm	MAE
1	100, 75	210.5, 174.65	82.54	100.47	101.32
2	75, 50	174.65, 138.76	38.7	53.30	59.80
3	50, 25	138.76, 102.88	13.55	24.94	34.39
4	25, 0	102.88, 66.99	0.56	0.74	0.84
5	0, −25	66.99, 65.61	0.29	0.44	0.71
6	25, −50	65.61, 64.23	0.23	0.67	1.35
7	50, −75	64.23, 62.85	0.23	0.41	0.44
8	75, −100	62.85, 61.47	0.21	0.28	0.45