Presentation Information

[22p-52A-16]Automatic identification of crystalline phases with Bayesian estimation in XRD

〇Ryo Murakami1, Yoshitaka Matsushita1, Kenji Nagata1, Shouno Hayaru2, Hideki Yoshikawa1 (1.NIMS, 2.UEC)

Keywords:

spectral decomposition,data-driven science

Our method successfully identifies the true crystal structure with high probability from a given set of 50 crystal structure candidates, and our method is also able to estimate the mixing ratio of each automatically deduced crystal phase with high accuracy. Our method can automatically identify each crystal phase with a confidence interval from XRD data of the whole observed area.