Modeling the Effect of Crude Oil Impacted Sand on the Properties of Concrete Using Artificial Neural Networks
Table 2
Details of various architectures examined.
N
ARCH
L.R
M
Activation funct.
R.M.S.E TR.
R.M.S.E TE.
IT
N1
2-2-1
0.01000
0.8
SIG-SIG-SIG
0.054328
0.033530
10000
N2
2-2-1
0.01000
0.8
TANH-TANH-TANH
0.029904
0.024476
10000
N3
2-2-1
0.01000
0.8
SIG-TANH-SIG
0.034249
0.025659
10000
N4
2-2-1
0.01000
0.8
TANH-SIG-TANH
0.022480
0.020762
10000
N5
2-2-1
0.05513
0.8
TANH-TANH-TANH
0.021212
0.020793
10000
N6
2-2-1
0.11000
0.8
TANH-TANH-TANH
0.019380
0.021277
20000
N7
2-2-1
0.10450
0.8
SIG-SIG-SIG
0.023245
0.021244
20000
N8
2-3-1
0.09928
0.8
TANH-TANH-TANH
0.016828
0.021612
20000
N9
2-3-1
0.09431
0.8
SIG-SIG-SIG
0.022184
0.022319
20000
N10
2-3-1
0.09928
0.8
SIG-SIG-SIG
0.017338
0.021098
20000
N11
2-3-1
0.09929
0.8
TANH-TANH-TANH
0.015696
0.027982
20000
N12
2-4-1
0.06464
0.8
TANH-TANH-TANH
0.014510
0.026469
20000
N13
2-4-1
0.06141
0.8
SIG-SIG-SIG
0.015890
0.026524
20000
N14
2-5-1
0.09500
0.6
TANH-TANH-TANH
0.014924
0.026600
20000
N15
2-6-1
0.09500
0.6
SIG-SIG-SIG
0.0199275
0.020644
20000
N16
2-6-1
0.06500
0.6
TANH-TANH-TANH
0.012363
0.027681
20000
N17
2-2-2-1
0.03000
0.6
TANH-TANH-TANH
0.021810
0.020928
20000
N18
2-2-2-1
0.95000
0.6
SIG-SIG-SIG
0.021958
0.021007
20000
N19
2-2-4-1
0.03033
0.6
TANH-TANH-TANH
0.020828
0.020437
20000
N20
2-2-1
0.03106
0.6
TANH-TANH-TANH
0.016831
0.036802
20000
N: network number; ARCH: network architecture; L.R: learning rate; M: momentum. R.M.S.E TR.: root mean square error for training data. R.M.S.E TE.: root mean square error for testing data. IT: maximum number of iterations.
