Table 9 Results of the effect of different transfer functions between 2 layers in predicting compressive strength of 7 and 28 days of self-compacting concrete.

From: Predicting the compressive strength of self-compacting concrete by developed African vulture optimization algorithm-Elman neural networks

 

Transfer functions

Tansig–Tansig

Tansig–logsig

Tansig–purline

Purline–logsig

Purline–purline

Purline–Tansig

Logsig–logsig

Logsig–purline

Logsig–tansig

MSE Test

\({F}_{c}7\)

11.69

168.14

15.31

167.82

533.25

175.10

166.84

21.83

15.10

 

\({F}_{c}28\)

18.59

253.05

31.43

338.63

2.05E + 3

45.89

253.65

41.39

83.43

MSE Train

\({F}_{c}7\)

12.57

135.80

17.27

134.10

265.59

19.36

17.86

203.78

134.54

 

\({F}_{c}28\)

18.59

223.79

28.22

337.86

1.32E + 3

55.50

222.97

44.53

81.35

R Test

\({F}_{c}7\)

0.96

0.41

0.95

0.44

0.30

0.39

0.43

0.92

0.95

 

\({F}_{c}28\)

0.97

0.67

0.95

6.2E−28

0.26

0.92

0.67

0.93

0.86

R Train

\({F}_{c}7\)

0.95

0.47

0.93

0.48

0.55

0.34

0.48

0.93

0.93

 

\({F}_{c}28\)

0.97

0.70

0.95

−1.7E−26

0.40

0.91

0.70

0.93

0.86