Presentation Information
[8p-P11-2]Machine Learning of HOMO and LUMO Energies Based on Differences Between Similar Molecules
〇(B)Hiroto Otokozawa1, Ren Sasaki1, Tomoharu Okada1, Yuki Mochizuki1, Hiroyuki Matsui1 (1.ROEL, Yamagata Univ.)
Keywords:
Materials Informatics,HOMO/LUMO Energy,Quantum Chemical Calculations
In recent years, materials informatics utilizing machine learning has significantly contributed to improving the efficiency of materials development. Our laboratory has developed a molecular design application, YU canvas, which enables rapid prediction of HOMO and LUMO energy levels computed by density functional theory (DFT) using machine learning. However, achieving high prediction accuracy remains a challenge. In this study, we aim to construct a more accurate prediction model by leveraging energy differences from structurally similar molecules with known DFT values.