Preferred attachment in affiliation networks (Q743425)

The authors study preferred attachment in growing affiliation networks, where vertices are linked to items independently at random, and the probability of an edge between the new vertex and an item \(w\) is proportional to the number of vertices already linked to \(w\). The asymptotic degree distribution of the evolving network is identified. It is shown that the fraction of vertices having degree \(i\) scales as \(i^{-2-\alpha}\ln i\) for large \(i\), where \(\alpha\) is a model parameter involving the intensity of the flow of customers (vertices) and the intensity of the newly arriving items.