jrSiCKLSNMF (version 1.2.3)

DetermineClusters: Perform clustering diagnostics

Description

A wrapper for the clValid and fviz_nbclust functions to perform clustering diagnostics

Usage

DetermineClusters(
  SickleJr,
  numclusts = 2:20,
  clusteringmethod = "kmeans",
  diagnosticmethods = c("wss", "silhouette", "gap_stat"),
  clValidvalidation = "internal",
  createDiagnosticplots = TRUE,
  runclValidDiagnostics = TRUE,
  printPlots = TRUE,
  printclValid = TRUE,
  subset = FALSE,
  subsetsize = 1000,
  seed = NULL
)

Value

An object of class SickleJr with cluster diagnostics added to its clusterdiagnostics slot

Arguments

SickleJr

An object of class SickleJr

numclusts

A vector of integers indicating the number of clusters to test

clusteringmethod

String holding the clustering method: defaults to k-means; since the other methods are not implemented in jrSiCKLSNMF, it is recommended to use k-means.

diagnosticmethods

Vector of strings indicating which methods to plot. Defaults to all three of the available: wss, silhouette, and gap_stat

clValidvalidation

String containing validation method to use for clValid. Defaults to internal.

createDiagnosticplots

Boolean indicating whether to create diagnostic plots for cluster size

runclValidDiagnostics

Boolean indicating whether to calculate the diagnostics from clValid

printPlots

Boolean indicating whether to print the diagnostic plots

printclValid

Boolean indicating whether to print the diagnostic results from clValid

subset

Boolean indicating whether to calculate the diagnostics on a subset of the data rather than on the whole dataset.

subsetsize

Numeric value indicating size of the subset

seed

Numeric value holding the random seed

References

clValidjrSiCKLSNMF

factoextrajrSiCKLSNMF

Examples

#Since these data are too small, the clValid diagnostics do not run
#properly. See the vignette for an example with the clValid diagnostics
SimSickleJrSmall<-DetermineClusters(SimSickleJrSmall,numclusts=2:5,runclValidDiagnostics=FALSE)