Compare clustering methods

Compare clustering methods

Usage

cmp_make_clusters(
  mat,
  method = setdiff(all_clustering_methods(), "mclust"),
  verbose = TRUE
)

cmp_make_plot(mat, clt, plot_type = c("mixed", "heatmap"), nrow = 3)

compare_clustering_methods(
  mat,
  method = setdiff(all_clustering_methods(), "mclust"),
  plot_type = c("mixed", "heatmap"),
  nrow = 3,
  verbose = TRUE
)

Arguments

mat

The similarity matrix.

method

Which methods to compare. All available methods are in all_clustering_methods(). A value of "all" takes all available methods. By default "mclust" is excluded because its long runtime.

verbose

Whether to print messages.

Ddetails The function compares following default clustering methods by default:

-kmeans see cluster_by_kmeans. -pam see cluster_by_pam. -dynamicTreeCut see cluster_by_dynamicTreeCut. -mclust see cluster_by_mclust. By default it is not included. -apcluster see cluster_by_apcluster. -hdbscan see cluster_by_hdbscan. -fast_greedy see cluster_by_fast_greedy. -louvain see cluster_by_louvain. -walktrap see cluster_by_walktrap. -MCL see cluster_by_MCL. -binary_cut see binary_cut.

Also the user-defined methods in all_clustering_methods are also compared.

clt

A list of clusterings from cmp_make_clusters().

plot_type

What type of plots to make. See Details.

nrow

Number of rows of the layout when plot_type is set to "heatmap".

Value

cmp_make_clusters() returns a list of cluster label vectors from different clustering methods.

cmp_make_plot() returns no value.

compare_clustering_methods() returns no value.

Details

For cmp_make_plot(), if plot_type is the default value "mixed", a figure with three panels will be generated:

• A heatmap of the similarity matrix with different classifications as row annotations.

• A heatmap of the pair-wise concordance of the classifications of every two clustering methods.

• Barplots of the difference scores for each method (calculated by difference_score), the number of clusters (total clusters and the clusters with size >= 5) and the mean similarity of the terms that are in the same clusters.

If plot_type is "heatmap". There are heatmaps for the similarity matrix under clusterings from different methods. The last panel is a table with the number of clusters under different clusterings.

compare_clustering_methods() is basically a wrapper function of cmp_make_clusters() and cmp_make_plot().

Examples

# \donttest{
mat = readRDS(system.file("extdata", "random_GO_BP_sim_mat.rds",
    package = "simplifyEnrichment"))
compare_clustering_methods(mat)
#> Cluster 500 terms by 'kmeans'...
#>  16 clusters, used 3.504993 secs.
#> Cluster 500 terms by 'pam'...
#>  5 clusters, used 21.06912 secs.
#> Cluster 500 terms by 'dynamicTreeCut'...
#>  60 clusters, used 0.577157 secs.
#> Cluster 500 terms by 'apcluster'...
#>  41 clusters, used 0.6789331 secs.
#> Cluster 500 terms by 'hdbscan'...
#>  14 clusters, used 0.172466 secs.
#> Cluster 500 terms by 'fast_greedy'...
#>  6 clusters, used 0.104732 secs.
#> Cluster 500 terms by 'louvain'...
#>  6 clusters, used 0.08450413 secs.
#> Cluster 500 terms by 'walktrap'...
#>  6 clusters, used 0.263876 secs.
#> Cluster 500 terms by 'MCL'...
#>  6 clusters, used 1.821683 secs.
#> Cluster 500 terms by 'binary_cut'...
#>  19 clusters, used 4.193117 secs.

compare_clustering_methods(mat, plot_type = "heatmap")
#> Cluster 500 terms by 'kmeans'...
#>  19 clusters, used 3.305151 secs.
#> Cluster 500 terms by 'pam'...
#>  5 clusters, used 21.00071 secs.
#> Cluster 500 terms by 'dynamicTreeCut'...
#>  60 clusters, used 0.1564171 secs.
#> Cluster 500 terms by 'apcluster'...
#>  41 clusters, used 0.7714469 secs.
#> Cluster 500 terms by 'hdbscan'...
#>  14 clusters, used 0.1589549 secs.
#> Cluster 500 terms by 'fast_greedy'...
#>  6 clusters, used 0.09291291 secs.
#> Cluster 500 terms by 'louvain'...
#>  6 clusters, used 0.07026696 secs.
#> Cluster 500 terms by 'walktrap'...
#>  6 clusters, used 0.2535839 secs.
#> Cluster 500 terms by 'MCL'...
#>  6 clusters, used 1.781449 secs.
#> Cluster 500 terms by 'binary_cut'...
#>  18 clusters, used 4.113788 secs.

# }