Framework for Converting Mechanistic Network Models to Probabilistic Models

Abstract: There are two prominent paradigms to the modeling of networks: in the first, referred to as the mechanistic approach, one specifies a set of domain-specific mechanistic rules that are used to grow or evolve the network over time; in the second, referred to as the probabilistic approach, one describes a model that specifies the likelihood of observing a given network. Mechanistic models are scalable and, in select cases, allow for analytical solutions for some of their properties, whereas probabilistic models have inferential tools available. Mechanistic models are appealing because they capture scientific processes that are hypothesized to be responsible for network generation. We introduce a generic framework for converting a mechanistic network model to a probabilistic network model. The proposed framework makes it possible to identify the essential network properties and their joint probability distribution for mechanistic network models, which enables addressing questions such as whether two mechanistic models generate networks with identical distributions of properties of interest, or whether a network property, such as clustering, is over- or under- represented in the generated networks compared to a reference model. The proposed framework is intended to bridge some of the gap that currently exists between mechanistic and probabilistic network models.