Molecular fingerprint design method for krypton/xenon separation metal organic framework membrane

The invention provides a molecular fingerprint design method for a krypton/xenon separation metal organic framework membrane, which comprises the following steps of: acquiring a descriptor of the metal organic framework membrane and a krypton/xenon separation performance parameter of an MOF (Metal Organic Framework) membrane, and establishing a data set based on the descriptor of the metal organic framework membrane and the krypton/xenon separation performance parameter of the MOF membrane; sorting the data sets according to the size of the permeation amount; selecting a machine learning algorithm to predict classification of the data set; the composition of the structure of the MOFs material is extracted from the data set in the algorithm with the highest prediction accuracy; the type of the molecular fingerprint is selected, and the composition of the structure of the extracted MOFs material is used for inputting the molecular fingerprint; and according to the classification prediction effect, defining that