Screening of New Additives of Active-Carbon-Supported Heteropoly Acid Catalyst for Friedel–Crafts Reaction by Gaussian Process Regression

Activity of heteropoly acid catalyst supported on active carbon for Friedel–Crafts reaction was predicted by Gaussian process regression (GPR), using the five main principal components of physicochemical properties of elements of the additives. Effective additives that promote the activity were predicted by the regression model and verified by experiments. The performance and accuracy of the regression model was increased using the expected improvement, which can suggest the additional experiments necessary for the improvement of the regression model. The performance of the regression model by GPR was superior to that of the radial basis function network (RBFN) or the support vector machine (SVM). In addition to the results by RBFN and SVM, an excellent effect of Pt addition was discovered by GPR.