Table 6 Prediction model performance on cross–project database (Scenario 2).

From: Machine learning approach for prediction of TBM performance and risk of jamming in Himalayan geology using a cross-project tunnelling database

Model

Training Set

Testing Set

R2

MAE

RMSE

MAPE

R2

MAE

RMSE

MAPE

XGBoost

0.974

1.426

2.424

2.220

0.972

1.708

2.919

3.330

RF

0.983

1.156

2.00

1.860

0.895

3.569

5.683

6.400

Bagging

0.978

1.189

2.252

1.870

0.986

1.288

2.108

2.210

Stacking

0.961

1.648

2.995

2.600

0.978

1.585

2.610

2.930

ANN

0.982

1.233

2.341

3.160

0.933

1.755

3.928

2.750

Back to article page