Fig 1.
Technology roadmap.
Fig 2.
Noise reduction by the sliding average algorithm.
Fig 3.
Wavelet noise reduction.
Fig 4.
Effect of noise reduction after two rounds.
Fig 5.
Results of curve fitting for different sizes.
Fig 6.
Schematic diagram of the LSTM structure.
Table 1.
Details of the SARIMA parameters.
Fig 7.
Schematic diagram of the SARIMA algorithm flow.
Fig 8.
Three stages of peristaltic deformation.
Fig 9.
T–t curve.
Table 2.
Improve the range of definition of tangent angle.
Fig 10.
Rock slope site photos.
Fig 11.
Engineering geological profile.
Fig 12.
Schematic diagram of the distribution of major fracture structures in the mine area and the location of the radar arrangement.
Fig 13.
Radar displacement monitoring data.
Fig 14.
Sliding average filtering global denoising.
Fig 15.
Wavelet localized region focused denoising.
Fig 16.
Results of separating the trend and periodic terms of displacement.
Fig 17.
Future time step prediction results of the trend term of displacement.
Fig 18.
Gray correlation coefficient calculation results.
Fig 19.
Prediction results of the future time step of the periodic term of displacement.
Fig 20.
Error line diagrams for LSTM–SARIMA hybrid models.
Fig 21.
Error line diagrams for RNN.
Fig 22.
Error line diagrams for SVM.
Fig 23.
Error line diagrams for XGBoost.
Fig 24.
Error line diagrams for LightGBM.
Fig 25.
Error line diagrams for Random forest.
Table 3.
Results of the four data-driven models.
Fig 26.
T–t curve for open pit displacement monitoring.
Fig 27.
Landslide site and landslide boundary diagram of open pit research area.