fig3
Figure 3. Predictive performance of the best model for each mechanical property across the PI-H and PI-F datasets: (A) ε for the PI-H dataset using RFR; (B) σ for the PI-H dataset using the GBR; (C) Et for the PI-H dataset using GBR; (D) ε for the PI-F dataset using GBR; (E) σ for the PI-F dataset using GBR; (F) Et for the PI-F dataset using GBR. RFR: Random forest regression; GBR: gradient boosting regression.