Estimating a graph from triad counts

An empirical graph [Ggrave] is described by a random graph model which generates [Ggrave] from an unknown graph G by independent removals and additions of edges. We consider the problem of estimating G by using the triad counts in [Ggrave], i.e. the numbers of different induced subgraphs of order three in [Ggrave]. We describe methods of estimating a transitive graph and a forest, and we indicate a possible approach for a general graph G.