Skip to main content
Figure 8 | Complex Adaptive Systems Modeling

Figure 8

From: Information theoretical methods for complex network structure reconstruction

Figure 8

Scientific collaboration network of researchers at INMEGEN after DPI pruning of indirect interactions - network 3. Panel A presents a SCN similar to Network 2 but indirect interactions have been pruned-out, (same as in Figure 7) after eliminating all indirect interactions by means of the application of the DPI. Network 3 is similar to Network 2, i.e. the effect of eliminating indirect interactions has a small impact. In many triadic interactions (triangles in the network) two-out of three interactions presented the same value so that no edge was eliminated. Panel B depicts the same network, nodes are size and color-coded according to their connectivity degree. The presence of a couple of well-connected individuals (big red nodes) as well as a number of medium-high connected ones (orange mid-sized nodes) points out to some kind of hierarchic structure. Research groups are easily identified, yet a certain degree of collaboration between most large groups is found. In panel C the connectivity degree distribution shows a power-law like behavior but with a smaller high-degree tail than in Network 2. In panel D the average clustering coefficient displays again a power-law like behavior. The reason for finite (non-zero) clustering coefficients is that a number of triangles remained in the network for two of its edges presented equal strength of interaction. Panel E presents the short path length distribution, which is unimodal but with a sharp expected separation length of 4. It is noticeable that eventhough DPI-pruning did not eliminated all the triadic connections in the network, the effect of network assessment to knock-out clear indirect interactions affected network navigability (average path length went from a clear value of three in Network 1, to between three and four in Network 2 and then to a clear value of three in Network 3).

Back to article page