Fig. 7: XGBoost model for CRSS prediction. | npj Computational Materials

Fig. 7: XGBoost model for CRSS prediction.

From: Unraveling dislocation-based strengthening in refractory multi-principal element alloys

Fig. 7

a Different feature combinations with their corresponding RMSE in CV tests, dashed lines show the variation of the smallest RMSE with different numbers of features and the best combination is marked by triangles. b, c Comparison between calculated and predicted CRSS/G for screw and edge dislocations, blue and red points representing train and test sets. d, e Corresponding feature importance. b, d for edge dislocation while c, e for screw.

Back to article page