Table 6 Predictive performance of the evaluated machine learning models for CMOD. Reported values are R2, RMSE, and VAF, each accompanied by 95% CIs obtained from 1000 bootstrap resamples.

From: Machine learning-based prediction of crack mouth opening displacement in ultra-high-performance concrete

Model

R2 (95% CI)

RMSE (95% CI)

VAF (95% CI)

DTR

0.62 (0.50–0.67)

0.14 (0.13–0.15)

0.62 (0.50–0.68)

SVR

0.87 (0.86–0.89)

0.08 (0.08–0.09)

0.87 (0.87–0.89)

NuSVR

0.88 (0.85–0.89)

0.08 (0.08–0.09)

0.88 (0.86–0.89)

GPR

0.89 (0.87–0.89)

0.08 (0.07–0.09)

0.88 (0.88–0.89)

XGBoost

0.71 (0.67–0.75)

0.12 (0.10–0.15)

0.71 (0.67–0.75)

RFR

0.78 (0.73–0.82)

0.11 (0.09–0.13)

0.78 (0.73–0.82)

GBR

0.80 (0.77–0.82)

0.10 (0.09–0.12)

0.80 (0.77–0.82)

ANN

0.85 (0.82–0.87)

0.09 (0.08–0.10)

0.86 (0.84–0.88)

TabPFN

0.90 (0.89–0.91)

0.07 (0.07–0.08)

0.90 (0.89–0.91)

Back to article page