Cluster similarity matrix by graph community detection methods

cluster_by_igraph(mat,
    method = c("cluster_fast_greedy",
    "cluster_leading_eigen",
    "cluster_louvain",
    "cluster_walktrap"),
    ...)

Arguments

mat: The similarity matrix.
method: The community detection method.
...: Other arguments passed to the corresponding community detection function, see Details.

Details

The symmetric similarity matrix can be treated as an adjacency matrix and constructed as a graph/network with the similarity values as the weight of hte edges. Thus, clustering the similarity matrix can be treated as detecting clusters/modules/communities from the graph.

Four methods implemented in igraph package can be used here:

cluster_fast_greedy: uses cluster_fast_greedy.
cluster_leading_eigen: uses cluster_leading_eigen.
cluster_louvain: uses cluster_louvain.
cluster_walktrap: uses cluster_walktrap.

Value

A vector of cluster labels (in numeric).

Examples

# There is no example
NULL
#> NULL