Prediction of the Frost Resistance of Iron Ore Tailings Concrete Based-BP Neural Network
Table 2
Predicted and experimental results of RDEM.
No.
Rr
NFTC
RDEM
Absolute error
Relative error (%)
No.
Rr
NFTC
RDEM
Absolute error
Relative error (%)
Exp
Pre
Exp
Pre
1
0.2
20
98.54
98.64
0.10
0.10
16
0.8
80
88.91
89.08
0.17
0.19
2
0.4
20
98.88
98.69
0.19
0.19
17
0
100
87.72
87.32
0.40
0.46
3
0.6
20
98.06
97.68
0.38
0.39
18
0.2
100
88.17
88.04
0.13
0.15
4
0.8
20
97.89
97.85
0.04
0.04
19
0.4
100
88.93
88.47
0.46
0.52
5
0
40
96.12
96.09
0.03
0.03
20
0.6
100
86.83
86.65
0.18
0.21
6
0.4
40
97
97.32
0.32
0.33
21
0
120
84.27
84.51
0.24
0.28
7
0.6
40
95.7
96.04
0.34
0.36
22
0.2
120
84.99
85.06
0.07
0.08
8
0.8
40
95.21
95.31
0.10
0.11
23
0.4
120
85.84
85.98
0.14
0.16
9
0
60
93.67
93.81
0.14
0.15
23
0.6
120
83.64
83.83
0.19
0.23
10
0.2
60
94.28
93.88
0.40
0.42
25
0.8
120
81.84
81.68
0.16
0.20
11
0.6
60
92.94
92.49
0.45
0.48
0
20
98.32
98.37
0.05
0.05
12
0.8
60
92.23
92.28
0.05
0.05
0.2
40
96.77
96.64
0.13
0.13
13
0
80
90.82
90.85
0.03
0.03
0.4
60
94.59
96.99
2.40
2.54
14
0.2
80
91.18
91.48
0.30
0.33
0.6
80
89.84
88.09
1.75
1.95
15
0.4
80
91.85
92.38
0.53
0.58
0.8
100
85.46
85.25
0.21
0.25
Note: E is the experimental value, P is the predictive value. The normal data for training and testing the BPNN. Marked with at the serial numbers are those for testing.
