Effective Augmentation Diversity Predicts Template Element Replacement Effectiveness in XRD-Based Crystal Phase Classification

This paper studies template element replacement as a data augmentation strategy for XRD-based crystal phase classification and shows that effective augmentation diversity predicts how well the replacement works.