IWA Publishing

Table 12

The results of the Kruskal–Wallis test used to determine if there were statistically significant differences between the applied predictive models

Output parameter	Applied models	p-value	Significantly different (95%)	Significantly different (99%)
L_j	RANFIS models	0.988	No	No
	ANFIS models	0.999	No	No
	ANFIS models, and RANFIS models	1	No	No
	ANFIS models, RANFIS models, and the empirical relation	<0.0001	Yes	Yes
h₂	ANFIS models, and the empirical relations	<0.0001	Yes	Yes
h₂	ANFIS models	1	No	No

Output parameter	Applied models	p-value	Significantly different (95%)	Significantly different (99%)
L_j	RANFIS models	0.988	No	No
	ANFIS models	0.999	No	No
	ANFIS models, and RANFIS models	1	No	No
	ANFIS models, RANFIS models, and the empirical relation	<0.0001	Yes	Yes
h₂	ANFIS models, and the empirical relations	<0.0001	Yes	Yes
h₂	ANFIS models	1	No	No