Fig. 3
From: Modelling sequences and temporal networks with dynamic community structures
Selection of US flight itineraries for a third-order model. The itineraries contain stops in Atlanta or Las Vegas. Edges incident on memories of the type x = (x t−1, Atlanta, x t−3) in red and x = (x t−1, LasVegas, x t−3) in blue. The node colours and overlaid hierarchical division derive from the n = 3 model inferred for the whole dataset
