Proposal of an Approach to Explore Photocatalysts by Two-stage Machine Learning Model

In the field of data-driven material development, bias in a dataset often causes difficulties in building a regression model when machine learning methods are applied. One of inorganic functional materials facing such a difficulty is photocatalysts. In this study, we propose a two-stage machine learning model to predict the activity for hydrogen evolution (H2/µmol h-1) from an aqueous solution containing sacrificial reagents over metal-sulfide photocatalysts under visible light irradiation. This two-stage machine learning model consists of the following two parts: a first regression model that predicts the activity for sacrificial hydrogen evolution and a second classification model that determines the reliability of the values predicted by the first regression model. We also propose a search scheme for variables related to the experimental conditions based on the proposed two-stage machine learning model. The proposed two-stage machine learning model improves the prediction accuracy of the activity compared with the first regression model.