A Cross-validation framework for several semi-supervised classification methods on graph nodes with Julia. It is possible to combine various Graph Similarities, Kernels and methods in order to classify nodes from graph data. Cross-validation is performed on a grid search with several parameters, depending on the method.
Most of the methods were used (and explained) in :
Guillaume Guex, Sylvain Courtain, Marco Saerens (2021) "Covariance and correlation measures on a graph in a generalized bag-of-paths formalism" in Journal of Complex Networks, Volume 8, Issue 6. https://academic.oup.com/comnet/article/8/6/cnaa026/6161491?login=true