An Assessment of a Proposed Hybrid Neural Network for Daily Flow Prediction in Arid Climate

Table 4

Optimum architecture of FF-MNN in Roodan watershed.

Input variable*	Hidden layer 1			Output layer
	Module	Transfer function	Learning rule	Transfer function	Learning rule
PCP(𝑡), PCP(𝑡−1), 𝑄(𝑡−1) = 3 cells	Module (1): 38 cells	Sigmoid	Levenberg-Marquardt	Linear: 1 cell	Momentum; (momentum = 0.8); (step size = 0.1)
	Module (2): 26 cells	Linear sigmoid

^∗PCP_(𝑡) is the present precipitation meanwhile PCP_(𝑡−1) is the precipitation for one day before present time. 𝑄_(𝑡−1) is the discharge for one day before present time.