Analysis of Updated Literature Data up to 2019 on the Oxidative Coupling of Methane Using an Extrapolative Machine?Learning Method to Identify Novel Catalysts

We have constructed and analyzed an updated dataset consisting of 4759 experimental datapoints for the oxidative coupling methane (OCM) reaction based on literature data reported before 2020 (?2019) using machine learning (ML) methods. Several ML methods, including random forest regression (RFR), extra trees (ETR), gradient boosting with XGBoost (XGB), were used in conjunction our proposed approach, which elemental features are as input representations rather than inputting catalyst compositions directly. A recent research trend, namely, extensive exploration Mn/Na2WO4/SiO2 systems years due to their high activity durability, was clearly reflected analysis. An model prediction outcome (C2 yield) successfully developed, feature importance scores SHapley Additive exPlanations (SHAP) values calculated ETR XGB, respectively, identify variables greatest influence performance observe how these important affect C2 yield OCM. The discovery optimization catalytic processes a “surrogate” explored, promising system candidates OCM identified. Notably, developed predicted catalysts containing elements that do not appear dataset. This demonstrates desirably potential enable extrapolative predictions ML-aided future catalysis research.