[20p-C601-1~16] 23.1 Joint Session N "Informatics"

4:45 PM - 5:00 PM JST(7:45 AM - 8:00 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[20p-C601-12] Adaptive Experimental Design for Model Selection and its Parameter Estimation

〇(M2)Tomohiro Nabika¹, Kenji Nagata², Shun Katakami¹, Mizumaki Masaichiro³, Masato Okada¹ (1.Univ. of Tokyo, 2.NIMS, 3.Kumamoto Univ.)

Keywords:

Experimental design,Bayesian inference,X-ray photon spectroscopy

Recently, experimental design using active learning has been proposed to improve the efficiency of experiments such as spectrum measurement. In this study, we constructed an active learning method suitable for mathematical model seclection and its parameter estimation based on Bayesian data analysis method. The proposed method can be applied to all measurements for which the candidate of mathematical model of the observation target is known, In this presentation, we apply the proposed method to artificial data of X-ray photoelectron spectroscopy and show the experimental time can be reduced to one-third of the conventional time.