Analysis of Complex Networks

16.5 Community Structure

Of great current interest is the identification of community groups, or modules, within networks. Stated informally, a community group is a portion of a network whose members are more tightly linked to one another than to other members of the network. A variety of approaches [3,11,19,20,22,26,27, 30,32] have been taken to explore this concept; see references [13,24] for useful reviews. Detecting community groups allows quantitative investigation of relevant subnetworks. Properties of the subnetworks may differ from the aggregate properties of the network as a whole, for example, modules on the World Wide Web are sets of topically related web pages. Thus, identification of community groups within a network is a first step toward understanding the heterogeneous substructures of the network.

Methods for identifying community groups can be categorized into distinct classes of networks, such as bipartite networks [5, 21]. This is immediately relevant for our study of FP networks, allowing us to examine the community structure in the bipartite networks. Communities are expected to be formed of groups of organizations engaged in similar R&D activities and the projects in which those organizations take part.

16.5.1 Modularity

To identify communities, we take as our starting point the modularity, introduced by [26]. Modularity makes intuitive notions of community groups precise by comparing network edges to those of a null model. The modularity Q is proportional ...