Consensus Community Detection

Runs a stochastic community detection algorithm multiple times and finds consensus communities via co-occurrence matrix thresholding. This approach produces more robust and stable community assignments than single 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
method: Community detection algorithm to use. Default "louvain". Must be a stochastic method (louvain, leiden, infomap, label_propagation, spinglass).
n_runs: Number of times to run the algorithm. Default 100.
threshold: Co-occurrence threshold for consensus. Default 0.5. Nodes that appear together in >= threshold proportion of runs are placed in the same community.
seed: Optional seed for reproducibility. If provided, the RNG state is initialized once before repeated runs and restored on exit.
...: Currently ignored. Each run calls the underlying igraph::cluster_*() function with its own defaults; no extra arguments are forwarded.

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 that of the final walktrap partition of the consensus graph, not of the original network.

Details

The algorithm works as follows:

Run the specified algorithm n_runs times using the current RNG stream
Build a co-occurrence matrix counting how often each pair of nodes appears in the same community
Normalize to proportions (0-1)
Threshold to create a consensus graph (edge if co-occurrence >= threshold)
Run walktrap on the consensus graph to get final communities

References

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

Examples

if (requireNamespace("igraph", quietly = TRUE)) {
  g <- igraph::make_graph("Zachary")

  # Consensus from 50 Louvain runs
  cc <- community_consensus(g, method = "louvain", n_runs = 50)
  print(cc)

  # Stricter threshold for more robust communities
  cc2 <- community_consensus(g, threshold = 0.7, n_runs = 100)
}
#> Community structure (consensus_louvain)
#>   Nodes: 34  | Communities: 4  | Modularity: 0.6409 
#>   Sizes: 12, 11, 5, 6 
#> 
#>  node community
#>     1         1
#>     2         1
#>     3         1
#>     4         1
#>     5         3
#>     6         3
#>     7         3
#>     8         1
#>     9         2
#>    10         1
#>    11         3
#>    12         1
#>    13         1
#>    14         1
#>    15         2
#>    16         2
#>    17         3
#>    18         1
#>    19         2
#>    20         1
#>    21         2
#>    22         1
#>    23         2
#>    24         4
#>    25         4
#>    26         4
#>    27         2
#>    28         4
#>    29         4
#>    30         2
#>    31         2
#>    32         4
#>    33         2
#>    34         2