Key points are not available for this paper at this time.
We present a stochastic model for the evolution of directed graphs over time. The infinitesimal transition rates for each arc in a directed graph depend only on the presence or absence of a reciprocated arc. The model reduces the entire directed graph to a set of~ independent and identically distributed dyad processeso The properties of the dyad process are discussed, four parameters of the stochastic model examined, and estimation strategies given based on maximum likelihood and the embeddability of the data for several sampling schemes.
Stanley Wasserman (Tue,) studied this question.