Add cluster number to a data frame

Clustering is either performed using stats::kmeans (method = "kmeans") on scaled numerical variables or using cluster::pam on a Gower dissimilarity matrix computed by cluster::daisy (method = "pam").

Usage

add_clusters(
  data,
  cols,
  newcol = NULL,
  k = 2,
  method = "kmeans",
  labels = TRUE,
  clean = TRUE
)

Arguments

data

A dataframe.

cols

A tidy selection of item columns.

newcol

Name of the new cluster column as a character vector. Set to NULL (default) to automatically build a name from the common column prefix, prefixed with "cls_".

k

Number of clusters to calculate. Set to NULL to output a scree plot for up to 10 clusters and automatically choose the number of clusters based on the elbow criterion.

method

The method as character value. One of "kmeans" (default) or "pam". For "kmeans" all items will be converted to numerical values and scaled using base::scale. Euclidean distance is used. For "pam", all items will be converted to categorical values by data_cat. A Gower dissimilarity matrix is used. Note that logical values are treated as asymmetrical, i.e. FALSE does not have a meaning, when computing the gower metric. Therefore cases with only FALSE values (no annotations, no codes) are removed.

labels

Whether to get the label of the cluster column from the common prefix of item column labels.

clean

Prepare data by data_clean.

Value

The input tibble with an additional cluster column (factor, prefixed "cls_"). The fit result is stored in the attribute stats.cluster.fit, the item names in stats.cluster.items, the scree-plot or silhouette-plot data in stats.cluster.diag and the method in stats.cluster.method.

Examples

library(volker)
ds <- volker::chatgpt

volker::add_clusters(ds, starts_with("cg_adoption"), k = 3)
#> # A tibble: 97 × 23
#>     case use_private use_work cg_adoption_advantage_01 cg_adoption_advantage_02
#>    <dbl>       <dbl>    <dbl>                    <dbl>                    <dbl>
#>  1   170           4        4                        3                        4
#>  2   183           1        1                        4                        3
#>  3   195           2        4                        5                        5
#>  4   212           5        5                        4                        4
#>  5   222           2        3                        3                        2
#>  6   236           3        1                        3                        2
#>  7   255           3        1                        3                        1
#>  8   297           3        4                        4                        3
#>  9   309           3        3                        3                        4
#> 10   325           2        1                        4                        1
#> # ℹ 87 more rows
#> # ℹ 18 more variables: cg_adoption_advantage_03 <dbl>,
#> #   cg_adoption_advantage_04 <dbl>, cg_adoption_fearofuse_01 <dbl>,
#> #   cg_adoption_fearofuse_02 <dbl>, cg_adoption_fearofuse_03 <dbl>,
#> #   cg_adoption_fearofuse_04 <dbl>, cg_adoption_social_01 <dbl>,
#> #   cg_adoption_social_02 <dbl>, cg_adoption_social_03 <dbl>,
#> #   cg_adoption_social_04 <dbl>, adopter <fct>, sd_age <dbl>, …
volker::add_clusters(ds, starts_with("cg_adoption"), k = 3, method = "pam")
#> # A tibble: 97 × 24
#>     case use_private use_work cg_adoption_advantage_01 cg_adoption_advantage_02
#>    <dbl>       <dbl>    <dbl> <fct>                    <fct>                   
#>  1   170           4        4 3                        4                       
#>  2   183           1        1 4                        3                       
#>  3   195           2        4 5                        5                       
#>  4   212           5        5 4                        4                       
#>  5   222           2        3 3                        2                       
#>  6   236           3        1 3                        2                       
#>  7   255           3        1 3                        1                       
#>  8   297           3        4 4                        3                       
#>  9   309           3        3 3                        4                       
#> 10   325           2        1 4                        1                       
#> # ℹ 87 more rows
#> # ℹ 19 more variables: cg_adoption_advantage_03 <fct>,
#> #   cg_adoption_advantage_04 <fct>, cg_adoption_fearofuse_01 <fct>,
#> #   cg_adoption_fearofuse_02 <fct>, cg_adoption_fearofuse_03 <fct>,
#> #   cg_adoption_fearofuse_04 <fct>, cg_adoption_social_01 <fct>,
#> #   cg_adoption_social_02 <fct>, cg_adoption_social_03 <fct>,
#> #   cg_adoption_social_04 <fct>, adopter <fct>, sd_age <dbl>, …