Another example

Download sample network file (15606 vertices, 45878 edges).

Draw network in 3D using one of the two possible approaches:

  • Layout/Pivot MDS/Random Pivots/3D by using different number of random pivots
  • Layout/Pivot MDS/Pivots from Cluster/3D by using community representatives as pivots
and play with different values of resolution parameter .

Resolution parameter r=1

For resolution r=1 we get the following correlations among partitions obtained by Louvain method and VOS Clustering (in both cases the obtained partition consists of 35 clusters):

Cramer's V: 0.80993631
Rajski(C1 <-> C2): 0.72959560
Rajski(C1 -> C2): 0.84392911
Rajski(C1 <- C2): 0.84339174
Adjusted Rand Index: 0.66048859

Best partition according to Louvain method, Q(r=1)=0.934149, 35 clusters

Best partition according to VOS Clustering, VOSQ(r=1)=0.936602, 35 clusters

Comparing Louvain method and VOS Clustering        Back to Pajek and Pajek-XXL Main page.