Constrained Community Detection in Social Networks

Community detection in networks is the process by which unusually well-connected sub-networks are identified–a central component of many applied network analyses. The paradigm modularity quality function optimization stipulates a partition network’s vertexes that maximizes difference between fraction edges within communities and corresponding expected if were randomly allocated among all vertex pairs while conserving degree distribution. incorporates exclusively topology has been extensively studied whereas integration constraints or external information on community composition largely remained unexplored. We define greedy, recursive-backtracking search procedure to identify constitution high-quality satisfy global constraint each be comprised at least one set so-called special apply our methodology identifying health care (HCCs) hospitals such HCC consists hospital wherein minimum number cardiac defibrillator surgeries performed. This restriction permits meaningful comparisons resulting standardizing distribution across network.