[9a-W321-1~11] 23.1 Joint Session N "Informatics"

[9a-W321-9] 【Highlight】Prediction system and feature importance for ferroelectricity of perovskite type oxides using first-principles calculation and machine learning

〇Junji Nishiyama¹, Takuya Takahashi¹, Hiroaki Takeda¹, Takaaki Tsurumi¹, Takuya Hoshina¹ (1.Tokyo Tech.)

Keywords:machine lerning, first-principles calculation, feroelectric

Acceleration of materials design by materials informatics are receiving a lot of attention in recent years. However, a few research on ferroelectrics has reported. In this study, database for machine learning of perovskite type oxides, which has typical ferroelectric, was prepared with first-principles calculation. Moreover, features were made for prediction of property and geometric parameters of ferroelectrics, and prediction system was built using the important features.