Fig. 6: Features analysis of the GFA prediction model after PCA. | npj Computational Materials

Fig. 6: Features analysis of the GFA prediction model after PCA.

From: A general and transferable deep learning framework for predicting phase formation in materials

Fig. 6

Projection of the feature vectors of 108 elements onto the plane spanned by the first and second principal axes. The percentage represents the ratio of the variance on the principal axis direction. Elements are colored according to their elemental groups. a Periodic table representation. b Randomized periodic table representation. The superscript 1–18 on element symbol represents the element’s group number; superscript 19 and 20 represent lanthanide and actinides, respectively.

Back to article page