16.3 Network Definition

Using the sysres EUPRO database, for each FP we construct a network containing the collaborative projects and all organizational subentities that are participants in those projects. An organization is linked to a project if and only if the organization is a member of the project. Since an edge never exists between two organizations or two projects, the network is bipartite. The network edges are unweighted; in principle, the edges could be assigned weights to reflect the strength of the participation, but the data needed to assign the network weights are not available.

We will also consider, for each FP, the projections of the bipartite networks onto unipartite networks of organizations and projects. The organization projections are constructed by taking the organizations as the vertices, with edges between any organizations that are at distance two in the corresponding bipartite network. Thus, organizations are neighbors in the projection network if they take part in one or more projects together. The project projections are similar, with project vertices linked when they have one or more participants in common. While the construction of the projection networks intrinsically loses information available in the bipartite networks, they can nonetheless be useful.

For the binary choice model, we construct another network using crosssection data on 191 organizations that are selected from the survey data. We employ the collaboration network of the respondents ...

Get Analysis of Complex Networks now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.