[17p-A401-1~15] 23.1 Joint Session N "Informatics"

[17p-A401-6] Feature Engineering of Ceramics Based on Conventional Knowledge in Material Chemistry and Prediction of Dielectric Constant of Paraelectric Materials

〇Noriaki Ozaki¹, Jun Ikeda¹ (1.Murata Mfg.)

Keywords:feature engineering, machine learning, dielectric materials

Acceralation of material research by machine learning (ML) has been essential for development in chemistry. However, input data for ML (features) proposed to date are not always apprehensible for experimental chemists, making it difficult to use or interpret ML. In this work, we developed features which are familiar and therefore highly usable for experimental chemists.