fig3

Uncertainty estimation affects predictor selection and its calibration improves materials optimization

Figure 3. Uncertainty estimation and calibration on the Ni-based dataset with different models. (A), (D), and (G) Miscalibration curves for SVR, XGBoost, and NN, respectively; (B), (E), and (H) Miscalibration curves after calibration; (C), (F), and (I) Relationship between RMSE and RMV before and after uncertainty calibration. SVR: Support vector regression; XGBoost: extreme gradient boosting; NN: neural network; RMSE: root-mean-square error; RMV: root mean variance.

Journal of Materials Informatics
ISSN 2770-372X (Online)
Follow Us

Portico

All published articles are preserved here permanently:

https://www.portico.org/publishers/oae/

Portico

All published articles are preserved here permanently:

https://www.portico.org/publishers/oae/