Community Detection with Self-Adapting Switching Based on Affinity

Figure 2

(a) We present the statistical results of for different values of µ from 0.1 to 0.8. The detailed information about the size and scale of network structure is shown in Section 3. The hollow box plot is calculated from the hub nodes’ neighbors in the same community, and the solid blue plot is calculated from the hub nodes’ neighbors in other communities. Each test is conducted by 100 times. (b) We statistically compute the normalization form for the neighbors inside and those outside the communities, respectively. Then, we get the mean values of for each hub node i, where the blue lines are derived from the inside neighbors of the hub node and the red lines are from outside. In both (a) and (b), the letter S denotes communities of size between 10 and 50 while the letter B denotes communities of size between 20 and 100.

(a)

(b)