Figure 5

We illustrate how the BiDir distance varies when using three different optimization functions, namely: Modularity, Infomap and stability. In (A) and (B) we illustrate network A and B, respectively, with node colouring corresponding to the communities detected. In (C) we show the BiDir distances obtained when comparing the modular structure of network A and B using each of the different optimization functions. A larger distance is obtained when using the modularity and Infomap function compared to the stability function. This is because the first two functions suffer from a ‘field-of-view’ limit and therefore over-partition network A. In general, the BiDir distance inherits the features of the optimization function deployed