Consensus Community Detection

Runs a stochastic community detection algorithm repeatedly and derives consensus communities by thresholding the co-occurrence matrix of the runs.

Usage

community_consensus(
  x,
  method = c("louvain", "leiden", "infomap", "label_propagation", "spinglass"),
  n_runs = 100,
  threshold = 0.5,
  seed = NULL,
  ...
)

com_consensus(
  x,
  method = c("louvain", "leiden", "infomap", "label_propagation", "spinglass"),
  n_runs = 100,
  threshold = 0.5,
  seed = NULL,
  ...
)

Arguments

x: Network input: matrix, igraph, network, cograph_network, or tna object. The louvain and leiden methods require an undirected network.
method: Community detection algorithm, one of "louvain" (default), "leiden", "infomap", "label_propagation" or "spinglass". The current code runs louvain for "spinglass".
n_runs: Number of runs. Default 100.
threshold: Co-occurrence threshold. Default 0.5. Pairs of nodes that share a community in at least this proportion of runs are linked in the consensus graph.
seed: Optional seed for reproducibility. If provided, the RNG state is initialized once before repeated runs and restored on exit.
...: Ignored. Each run calls the igraph function with its own defaults and without weights or resolution settings. For leiden this is the CPM objective with resolution 1.

Value

A cograph_communities data frame (columns node and community) holding the consensus membership. Its "algorithm" attribute is "consensus_<method>" and its "modularity" attribute is the modularity of the final walktrap partition computed on the consensus graph.

Details

The algorithm is run n_runs times on the current random number stream. The proportion of runs in which each pair of nodes shares a community forms the co-occurrence matrix. Pairs with a proportion of at least threshold are linked in an unweighted consensus graph, and walktrap on that graph gives the final communities.

References

Lancichinetti, A., & Fortunato, S. (2012). Consensus clustering in complex networks. Scientific Reports, 2, 336.

Examples

community_consensus(to_undirected(regulation_net), method = "louvain",
                    n_runs = 10, seed = 1)
#> Community structure (consensus_louvain)
#>   Nodes: 10  | Communities: 2  | Modularity: 0.5 
#>   Sizes: 5, 5 
#> 
#>        node community
#>     Explore         2
#>        Plan         1
#>     Monitor         1
#>       Adapt         2
#>     Reflect         2
#>     Discuss         2
#>  Synthesize         2
#>    Evaluate         1
#>      Create         1
#>       Share         1