Figure 3
The impact of (a), (b) broadening, (c), (d) baseline noise and (e), (f) multiple impurity phases on ML predictions for unit-cell lengths for (left-hand column) cubic and (right-hand column) triclinic crystal systems. NM refers to a pattern with no modifications and M refers to a pattern with the corresponding experimental modification. The notation A/B indicates training with modification A (NM or M) and testing with modification B (NM or M). The null column is a prediction based on the mean lattice parameters of the data set. The performance of ML models is greatly improved when training and testing with modifications (M/M) relative to training on unmodified data and testing on modified data (NM/M). |