The following tutorials present several random graph models, such as Erdos-Renyi model, stochastic block model, and random dot product graph (RDPG) model. These models provide a basis for studying random graphs.
Inference on random graphs depends on low-dimensional Euclidean representation of the vertices of graphs, known as graph embeddings, typically given by spectral decompositions of adjacency or Laplacian matrices. Below are tutorials for computing graph embeddings of single graph and multiple graphs.
Statistical testing on graphs requires specialized methodology in order to account for the fact that the edges and nodes of a graph are dependent on one another. Below are tutorials for robust statistical hypothesis testing on multiple graphs.