Fig. 5: Improvement in hardness prediction with newly generated features. | npj Computational Materials

Fig. 5: Improvement in hardness prediction with newly generated features.

From: Elemental numerical descriptions to enhance classification and regression model performance for high-entropy alloys

Fig. 5

Comparison of a R2 and b MAE values for machine learning models based on material feature subsets that achieve the lowest test error, with and without the inclusion of our newly generated features. The statistical significance of the improvement is indicated by the t test P value (<0.01).

Back to article page