Skip to main content
Fig. 13 | Journal of Cheminformatics

Fig. 13

From: MOFGalaxyNet: a social network analysis for predicting guest accessibility in metal–organic frameworks utilizing graph convolutional networks

Fig. 13

The process of predicting guest accessibility for new MOFs. To predict pore size for MOFs not present in MOFGalaxyNet, the adjacency matrix must be updated by adding the new MOFs to the graph. Relevant MOF information with the same data structure as training data must be extracted and vectorized for similarity measurement prior to updating the matrix. The GCN model is then fed with the updated matrix and new MOFs to make predictions about PLD size

Back to article page