Cluster Matrix of Binary Censored Data

Performs clustering of a matrix of 0s and 1s, ie. the censoring indicator columns for multiple variables. If there are multiple detection limits within a column first convert the 0/1 designated to be Below vs Above the highest detection limit in the column. Detection limits may differ among columns.

binaryClust(
  data,
  method = "ward.D2",
  group = NULL,
  ncluster = NULL,
  plotncluster = TRUE,
  clustIndex = "all"
)

Arguments

data

A data frame containing only the 0/1 columns, one column per chemical parameter.

method

Method of forming clusters. The default is "ward.D2", which is appropriate for a variety of types of data. Another appropriate option is “average” – average distances between cluster centers. See the vegan package for other possible clustering methods.

group

Optional grouping variable. If used, sites being clustered will be represented by their group name, rather than by the row number.

ncluster

Optional number of clusters to be differentiated on the graph. Clusters are fenced off with rectangles.

plotncluster

default is TRUE logical flags to add identification of clusters on dendrogram

clustIndex

Optional, if not specified, potential number of clusters will be determined based on the mean best number of clusters across all indicies. For a specific index, see details

Value

Prints a cluster dendrogram based on clustering method and outputs a list of clusters and hierarchical cluster analysis results

Details

If a specific index is desired to determine the best number of clusters see NbClust::NbClust for index values.

References

Helsel, D.R., 2011. Statistics for Censored Environmental Data using Minitab and R, 2nd ed. John Wiley & Sons, USA, N.J.

Examples

data(PbHeron)

# without group specified
binaryClust(PbHeron[,4:15])
#> Warning: NaNs produced

#> $hclust
#> 
#> Call:
#> hclust(d = simple_matching_coeff, method = method)
#> 
#> Cluster method   : ward.D2 
#> Distance         : simplematch 
#> Number of objects: 27 
#> 
#> 
#> $clusters
#>  [1] 1 1 2 1 2 3 4 1 1 3 1 1 2 2 3 3 1 3 3 3 3 3 1 3 1 3 3
#> 

# With Group argument
binaryClust(PbHeron[,4:15],group=PbHeron$DosageGroup)
#> Warning: NaNs produced

#> $hclust
#> 
#> Call:
#> hclust(d = simple_matching_coeff, method = method)
#> 
#> Cluster method   : ward.D2 
#> Distance         : simplematch 
#> Number of objects: 27 
#> 
#> 
#> $clusters
#>  [1] 1 1 2 1 2 3 4 1 1 3 1 1 2 2 3 3 1 3 3 3 3 3 1 3 1 3 3
#>