[19p-C601-1~15] 23.1 Joint Session N "Informatics"

2:15 PM - 2:30 PM JST(5:15 AM - 5:30 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[19p-C601-4] Extraction of polymer refractive index laws
by symbolic regression using bayesian information criterion

〇Naoki Yamane¹, Kan Hatakeyama², Yuma Iwasaki³, Yasuhiko Igarash¹ (1.Univ. of Tsukuba, 2.Tokyo Inst. of Tech., 3.NIMS)

Keywords:

Neural network potential,Symbolic regression,Bayesian information criterion

Neural potentials are deep learning models that rapidly approximate physicochemical calculations on the atomic scale, and are rapidly gaining attention in recent years, but they are computed by general laws and involve systematic errors. In this study, a symbolic regression is used to find a function f that minimises the Bayesian information criterion for the predicted value and the objective variable, and a correction formula is obtained. The results of the analysis show that the relational equation can be obtained by analysing real data of organic molecular materials with a high refractive index.