Abstract
Temporal networks are widely used models for describing the architecture of complex systems. Network memory -- that is the dependence of a temporal network's structure on its past -- has been shown to play a prominent role in diffusion, epidemics and other processes occurring over the network, and even to alter its community structure. Recent works have proposed to estimate the length of memory in a temporal network by using high-order Markov models. Here we show that network memory is inherently multidimensional and cannot be meaningfully reduced to a single scalar quantity. Accordingly, we introduce a mathematical framework for defining and efficiently estimating the microscopic shape of memory, which fully characterises how the activity of each link intertwines with the activities of all other links. We validate our methodology on a wide range of synthetic models of temporal networks with tuneable memory, and subsequently study the heterogeneous shapes of memory emerging in various real-world networks.
Abstract (translated)
URL
https://arxiv.org/abs/2004.12784
