daltoolbox (version 1.3.767)

cluster_louvain_graph: Louvain community detection

Description

Graph community detection using igraph::cluster_louvain.

Usage

cluster_louvain_graph(weights = NULL)

Value

returns a Louvain clustering object.

Arguments

weights

optional edge weights to pass to cluster_louvain

Details

Accepts an igraph object and returns community memberships. Requires the igraph package.

The base clusterer() uses wcss and entropy as generic defaults, but cluster_louvain_graph() specializes both because graph community detection is not a point-to-centroid clustering problem.

Default evaluation in cluster_louvain_graph() is:

• main metric: metric_modularity()

• internal evaluation: modularity

• external evaluation: none

The general wcss fallback is not used because graph communities are judged by network connectivity structure, not by geometric dispersion. The generic external entropy default is also not kept because this class operates on an igraph object and has no natural reference labels unless the user supplies an external comparison separately.

References

Blondel, V. D., Guillaume, J.-L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. J. Statistical Mechanics. Newman, M. E. J. (2006). Modularity and community structure in networks.

Examples

if (requireNamespace("igraph", quietly = TRUE)) {
 g <- igraph::sample_gnp(n = 20, p = 0.15)
 model <- cluster_louvain_graph()
 model <- fit(model, g)
 clu <- cluster(model, g)
 table(clu)
}