Page 252 - Read Online

Liu et al. J. Mater. Inf. 2026, 6, 18 Page 17 of 29

Figure 8. A performance comparison chart of 12 machine learning model test sets constructed based on the optimal strength feature
subset. ML: Machine learning; R : the coefficient of determination; RMSE: root mean square error; Bagging: bootstrap aggregating; RF:
2
random forest; GBRT: gradient boosting regression tree; ExtraTree: extremely randomized tree; AdaBoost: adaptive boosting; LGBM: light
gradient boosting machine; XGBoost: eXtreme gradient boosting; KNN: K-nearest neighbor; DT: decision tree; ANN: artificial neural
network; SVR: support vector regression.

Figure 9. Four-parameter hyperparameter optimization was performed on the random forest model fitted to the strength dataset. (A)
Maximum depth; (B) Minimum sample segmentation; (C) Minimum sample leaf; (D) n_estimators; (E) Strength prediction graph of the
best RF model under the optimal combination of hyperparameters. RF: Random forest; R : the coefficient of determination; RMSE: root
2
mean square error.

dataset, and an R of 0.82 with an RMSE of 3.66% on the test dataset, confirming the robustness and high
2
applicability of RF for predicting the ductility behavior of titanium alloys.

To verify model robustness and rule out potential overfitting, learning curves for both strength (GBRT) and