[19a-Z32-1~10] 23.1 Joint Session N "Informatics"

△ [19a-Z32-1] Bayesian Estimation for Multi-element XPS Spectrum Analysis using Basis Functions with Multiple Peaks

〇Atsushi Machida¹, Kenji Nagata², Ryo Murakami³, Hiroshi Shinotsuka², Hayaru Shouno³, Hideki Yoshikawa², Masato Okada^1,2 (1.Univ. of Tokyo, 2.NIMS, 3.UEC)

Keywords:data-driven science, Bayesian statistics

XPS is a widely used measurement technique for surface analysis of materials, but its analysis has a problem that the arbitrariness of the operator is included in the analysis results. In this study, we proposed a new Bayesian spectral decomposition method using exchange Monte Carlo and conducted experiments on artificial data. This method provides the posterior distribution of the model parameters, which allows us not only to estimate the unknown samples correctly, but also to discuss the confidence intervals of the estimated values.