抄録
In researching, developing, and manufacturing highly functional materials such as membrane materials, it becomes common to utilize chemical data and chemical engineering data for machine learning to improve the efficiency of molecular design, material design, process design, and process control. It is important to construct mathematical model y ＝f (x) with high predictive ability between explanatory variables x and objective variables y, and then, y values can be predicted from x values using the constructed model, and x values can be designed to meet target y values. In this article, as examples of research in chemoinformatics, materials informatics, and process informatics, the estimation of prediction errors in new samples, modeling of metal–organic frameworks with machine learning, adaptive design of experiments with direct inverse analysis for designs of molecules, materials, and processes, and prediction of future transmembrane pressure in a drinking water treatment process are introduced.
