Table 4 Comparison of evaluation indexes of measuring point S5 by four prediction models.

From: Monitoring and deformation of deep excavation engineering based on DFOS technology and hybrid deep learning

Prediction models

Determination coefficient \({\text{R}}^{2}\)

Root mean square error (RMSE)

Mean absolute error (MAE)

Mean bias error (MBE)

BP

0.8650

0.3824

0.3078

− 0.2976

LSTM

0.8772

0.3647

0.2749

− 0.1819

CNN–LSTM

0.9206

0.2932

0.2591

0.2395

CNN–LSTM–SAM

0.9724

0.1728

0.1267

− 0.0188

Back to article page