372 3.80 to 3.46 mm and MAE from 2.54 to 2.36 mm. The East changed only marginally, whereas performance worsened in the West and slightly in the Crest. KGE was not used for zone-wise interpretation because values remained low, with a maximum of 0.356 in the Crest. This indicates that, although absolute errors were acceptable, the models did not reproduce the temporal variability of the increment series sufficiently well for KGE to provide a robust basis for interpretation, increasing the risk of overfitting. Therefore, performance was assessed primarily using MAE and RMSE. Table 5 – Zone test performance for one-step-ahead forecasting of Δ4á Figure 4 – One-step-ahead forecasts of vertical deformation increments (Δ4á) from zone-specific LightGBM models versus MT-DInSAR observations with uncertainty envelopes. Figure 4 compares one-step-ahead LightGBM predictions of Δ4á with MT-DInSAR observations for representative points in each zone, together with the per-epoch observational uncertainty, 8,J. Across all four examples, the model reproduces the oscillatory behaviour of the increment series and captures most changes in sign, with point-level correlations ranging from 0.74 to 0.80 and MAE from 2.43 to 3.30 mm. Agreement is strongest for the Toe example (MAE = 2.43 mm; r = 0.78), while the East example shows the Meteorological data Metric Crest East Toe West No RSME 4.687 4.445 3.802 4.924 MAE 3.287 2.712 2.539 3.206 Yes RSME 4.712 4.422 3.646 5.352 MAE 3.132 2.980 2.360 3.372
RkJQdWJsaXNoZXIy MTM0Mzk2