Prediction of Chemical Reaction with Catalyst by Sparse Modeling

Abstract

In this talk, we introduce catalytic reaction prediction by sparse modeling as one of the practical cases of materials informatics. Specifically, we use sparse modeling to solve the regression problem, which uses the phosphoric acid catalyst used for the reaction in the epoxidation of olefins as input, and outputs the experimental results of the reaction yield. The information of charge and vibration mode are used as descriptors of phosphoric acid catalyst using first principle calculation. We introduce the result of the case of 14 experimental data.