First-Level Upperbound Screening
Assume that similarity searching is carried out to identify nearest neighbours that have similarities above some threshold similarity, T
Use of an initial calculation to check whether the clique-based similarity could possibly be > T; if not, then there is no need to invoke the clique-detection algorithm
The upperbounds are calculated using connectivity distributions that identify the maximum possible number of matching nodes and edges, and hence the size, of the largest possible MCSE