Artificial Neural Network - Water Quality Management: Assessment and Interpretation - page 78

Fig. 5.1 Schematic of an arti cial neuron

Hornik 1991 , 1993 ; Leshno et al. 1993 ). Figure 5.2 presents the schematic of the
ANN. Figure 5.3 shows schematic tangent

—

sigmoid transfer function and linear
transfer function. Number of neurons in hidden layers for each model may be
obtained using trial and error. In this network, number of input vector component.
R = the number of input vector components; S 1 and S 2 = number of Neurals in
hidden and output layers, respectively.
Number of network output would be S 2 . Function of the network can be mod-
eled by Eqs. 5.3 and 5.4 (Menhaj 1998 ):

( )

j ð t Þ = F X R

1 1

j ; i p i ð t Þþ b

1
j

a

i ¼ 1 w

ð 5 : 3 Þ

( )

j ð t Þ = G s1

2 2 1 2

k

a

j ¼ 1 w

k ; i a J ð t Þþ b

ð 5 : 4 Þ

Fig. 5.2 Transfer function
tangent sigmoid