Finding Statistically Significant Communities in Networks
Figure 16
Application of OSLOM to real networks: US airport network.
The maps show the position of the airports, which are represented by symbols, indicating the communities found by applying OSLOM directly to the corresponding network, without exploiting the information of previous snapshots. The diagram shows the “seasonality” of air traffic. The normalized mutual information (diamonds) was computed comparing the cover of the system at time adjusted by OSLOM on the network at time
, and the cover obtained by applying OSLOM directly to the system at time
. The circles are estimates of the similarity of the network matrices of snapshots separated by
(one year). For each year we took four snapshots, by cumulating the traffic of each trimester. The most stable networks are typically in winter (vertical lines).