Prediction of Seepage Pressure Based on Memory Cells and Significance Analysis of Influencing Factors
Table 1
Comparison of model prediction results.
Model
Absolute error (m)
Relative error (%)
Iteration time (s)
Iteration times (times)
R2
Maximum value
Minimum value
Median
Mean value
Maximum value
Minimum value
Median
Mean value
RNN model
0.109
0.004
0.031
0.040
15.84
0.59
4.85
6.08
10.85
775
0.89
BP model
0.156
0.011
0.051
0.064
22.55
1.67
7.83
9.85
15.55
1344
0.83
RS-LSTM model
0.054
0.000
0.015
0.020
9.54
0.03
2.34
3.00
6.37
278
0.97
Remarks: absolute error = |measured value of seepage pressure - predicted value of seepage pressure|; relative error = absolute error/water head × 100%.
