Application of machine learning to the teaching of membrane technology with the assistance of supercritical carbon dioxide
[Introduction]An innovative teaching mode combining the supercritical carbon dioxide(CO2)assisted preparation of metal-organic framework(MOF)membranes and machine learning technology was explored to enhance the undergraduates'abilities in experimental data processing,analysis,and prediction.[Method]During the teaching process,students use a data extraction tool firstly to obtain CO2 adsorption isotherms from complex experimental graphs,creating a dataset containing data points and material properties.Subsequently,students apply extreme gradient boosting(XGB)and random forest(RF)algorithms to train and test the dataset,establishing the predictive models for the properties of materials.[Result]The data extraction tool significantly improves the efficiency and accuracy of data processing.The XGB algorithm shows good accuracy and interpretability in predicting the material properties.Through this teaching mode,which combines data extraction and machine learning,students achieve significant progress in data processing and prediction,enhancing their data analysis and research capabilities.[Conclusion]The teaching mode proposed in this paper provides an effective solution for experimental data processing and analysis in the fields of chemical engineering and material science with high application value.