compare: Compare clustering methods

cmp_make_clustersR Documentation

Compare clustering methods

Description

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".

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().

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.

Examples


mat = readRDS(system.file("extdata", "random_GO_BP_sim_mat.rds",
    package = "simplifyEnrichment"))
compare_clustering_methods(mat)
compare_clustering_methods(mat, plot_type = "heatmap")


jokergoo/simplifyGO documentation built on Sept. 18, 2024, 9:45 p.m.