Table 4 Comparative performance of predictive models on burst pressure prediction.

From: physically interpretable residual strength prediction of corroded pipelines via symbolic Bayesian networks

Model

MSE

RMSE

MAE

\(R^2\)

Linear regression

25.710

5.071

3.664

0.531

Ridge

26.560

5.154

3.820

0.513

KNN

13.836

3.720

2.223

0.756

RF

9.079

3.013

1.690

0.875

SVR

30.816

5.551

3.747

0.479

GBDT

4.691

2.166

1.503

0.913

XGBoost

3.614

1.901

1.009

0.923

SyBN

1.700

1.304

0.968

0.966

Back to article page