[23p-P14-1~5] 23 Joint Session N "Informatics"(Poster)

[23p-P14-2] Improving the Accuracy of HOMO/LUMO Prediction of Organic Molecules by Machine Learning

〇(M1)Yuki Mochizuki¹, Tomoharu Okada¹, Hiroyuki Matsui¹ (1.ROEL, Yamagata Univ.)

Keywords:machine learning, organic molecule, properties prediction

In the past, the HOMO and LUMO levels of organic molecules were predicted by experimental measurements or theoretical calculations. These methods spend a lot of time and money, which are required the knowledge of physical chemistry and computational science. The solution to this subject is “Yamagata Chemical Canvas”, a tool that can rapidly predict the HOMO and LUMO levels with molecular formulas by machine learning. The tool was developed and published by Matsui lab. However, the tool still has a disadvantage of poor prediction accuracy. The purpose of this research is to improve the prediction of the HOMO and LUMO levels of organic molecules.