Table 1 Evaluation metrics of different models

From: Intelligent prediction of residual strength in blended hydrogen–natural gas pipelines with crack-in-dent defects

Model

R2

MSE

MAE

MAPE

Training

Test

Training

Test

Training

Test

Training

Test

TabPFN

0.9976

0.9961

0.0104

0.0178

0.0722

0.1020

0.0052

0.0077

LightGBM

0.9960

0.9859

0.0172

0.0639

0.0854

0.1665

0.0066

0.0129

XGBoost

0.9998

0.9810

0.0007

0.0863

0.0162

0.2117

0.0012

0.0155

RF

0.9881

0.9047

0.0514

0.4327

0.1719

0.4921

0.0126

0.0367

  1. Bold represents the data with the best performance.
Back to article page