ntlvn

Communities by the Louvain method

R

ntlvn(from, to, w = NULL, nodes = NULL, res = 1, seed = NULL)

Groups nodes to raise the modularity (see ntmod()) of the undirected, weighted graph (Blondel et al. 2008). Each level visits the nodes in node order when seed is NULL, or in a random order drawn from seed (one stream for the whole run), and moves each to the neighbouring community with the largest gain in modularity, the first met among equal gains, when that gain beats staying by more than 1e-12 of the node’s strength; passes repeat until none moves a node. The communities then become the nodes of the next level, and a level that moves nothing ends the run. The result is the same on every platform for the same input and seed.

Parameters

from, to

Labels of the two ends of each edge: vectors of the same length (zero when nodes gives every node) with no missing values.

w

NULL, or one positive weight per edge (a tie’s strength, such as the number of co-offences).

nodes

NULL, or every node’s label once: the node order of the results and the way to include isolated nodes.

res

The resolution: one positive number; above 1 favours smaller communities.

seed

NULL for node order, or one whole number from 0.

Returns

A list: memb, a data frame with node and comm (communities numbered 1, 2, … in the order of their first node); q, its modularity; levels, the levels that moved a node.

Notes

The method is a heuristic: different orders reach different local optima (on Zachary’s karate club, modularity from about 0.39 to the optimum 0.4198), so run several seeds and keep the best. Graph rules as in ntdeg(); isolated nodes stay alone.

References

Blondel, V. D., Guillaume, J.-L., Lambiotte, R. and Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of Statistical Mechanics, P10008. doi:10.1088/1742-5468/2008/10/P10008

Examples

e <- c(1, 1, 2, 4, 4, 5, 3)
f <- c(2, 3, 3, 5, 6, 6, 4)
ntlvn(e, f)

Generated from the R help page: ntlvn in the R reference.