doMIsaul (version 1.0.1)

MultiCons: MultiCons Consensus Clustering Algorithm

Description

Performs MultiCons clustering, from Al-Najdi et Al. For some reason, if you want to use mclust() clustering, the package needs to be loaded manually

Usage

MultiCons(
  DB,
  Clust_entry = FALSE,
  Clustering_selection = c("kmeans", "pam", "OPTICS", "agghc", "AGNES", "DIANA",
    "MCLUST", "CMeans", "FANNY", "BaggedClust"),
  num_algo = 10,
  maxClust = 10,
  sim.indice = "Jaccard",
  returnAll = FALSE,
  Plot = TRUE,
  verbose = FALSE
)

Arguments

DB

Either data or dataframe of partitions.

Clust_entry

Is DB partitions (TRUE) or data (FALSE).

Clustering_selection

If DB is data, clustering algorithm to select among. Must be included in default value.

num_algo

Number of clustering algorithms to perform.

maxClust

Maximum number of clusters.

sim.indice

Index for defining best partition. Passed to clusterCrit::extCriteria)(), see clusterCrit::getCriteriaNames(FALSE) for other available indexes. If more than one index are given, only the first one will be used.

returnAll

Should all partitions (TRUE) or only the best (FALSE) be returned.

Plot

Should tree be plotted.

verbose

Passed on to mclust() and other functions.

Value

A list of 2: performances and partitions. If returnAll is TRUE, both elements of the list contain results for all levels of the tree, else they only contain the results for the best level of the tree.

Examples

# NOT RUN {
library(mclust)
### With clustering algorithm choices
MultiCons(iris[, 1:4],
          Clustering_selection = c("kmeans", "pam", "DIANA", "MCLUST"),
          Plot = TRUE)
### With a manual clustering entry
parts <- data.frame(factor(rep(c(1,2,3), each = 50)),
                    factor(rep(c(1,2,3), times = c(100, 25, 25))),
                    factor(rep(c(1,2), times = c(50, 100))),
                    factor(rep(c(3, 2, 1), times = c(120, 10, 20))),
                    stringsAsFactors = TRUE)
MultiCons(parts, Clust_entry = TRUE, Plot = TRUE)
# }