how can I use 'silhouette_avg' with 'tune_cluster'?

[Ginseng7]

sorry for bad english.

Im trying to get 'k'for k-mean clutering.
I found method to get 'k' with elbow method by 'sse_ratio' like in later line.

#########################
library(workflows)
library(tune)

rec_spec <- recipe(~., data = mtcars) %>%
step_normalize(all_numeric_predictors()) %>%
step_pca(all_numeric_predictors())

kmeans_spec <- k_means(num_clusters = tune())

wflow <- workflow() %>%
add_recipe(rec_spec) %>%
add_model(kmeans_spec)

grid <- tibble(num_clusters = 1:10)

set.seed(4400)
folds <- vfold_cv(mtcars, v = 5)

res <- tune_cluster(
wflow,
resamples = folds,
grid = grid,
metrics = cluster_metric_set(sse_ratio) # want to get 'silhouette_avg'. how can i do that?
)

collect_metrics(res) → res_metrics

res_metrics %>%
filter(.metric == "sse_ratio") %>%
ggplot(aes(x = num_clusters, y = mean)) +
geom_point() +
geom_line() +
theme_minimal() +
ylab("mean WSS/TSS ratio, over 5 folds") +
xlab("Number of clusters") +
scale_x_continuous(breaks = 1:10)
#####################

I've tried to get 'k' with silhouette_avg but failed.

putting "silhouette_avg", "silhouette_avg(dists = dists)", "silhouette_ave, dists=dists"...
but failed.

how can i get table of silhouette instead of sse ratio?

thanks for your help.

[Emilhvitfeldt]

You can get the average silhouette values, by setting silhouette_avg in cluster_metric_set(). You might need to update your tidyclust version to the CRAN version

library(tidymodels)
library(tidyclust)

rec_spec <- recipe(~., data = mtcars) %>%
  step_normalize(all_numeric_predictors()) %>%
  step_pca(all_numeric_predictors())

kmeans_spec <- k_means(num_clusters = tune())

wflow <- workflow() %>%
  add_recipe(rec_spec) %>%
  add_model(kmeans_spec)

grid <- tibble(num_clusters = 1:10)

set.seed(4400)
folds <- vfold_cv(mtcars, v = 5)

res <- tune_cluster(
  wflow,
  resamples = folds,
  grid = grid,
  metrics = cluster_metric_set(silhouette_avg)
)

collect_metrics(res)
#> # A tibble: 10 × 7
#>    num_clusters .metric        .estimator    mean     n std_err .config         
#>           <int> <chr>          <chr>        <dbl> <int>   <dbl> <chr>           
#>  1            1 silhouette_avg standard   NaN         0 NA      Preprocessor1_M…
#>  2            2 silhouette_avg standard     0.427     5  0.0123 Preprocessor1_M…
#>  3            3 silhouette_avg standard     0.408     5  0.0181 Preprocessor1_M…
#>  4            4 silhouette_avg standard     0.400     5  0.0338 Preprocessor1_M…
#>  5            5 silhouette_avg standard     0.447     5  0.0156 Preprocessor1_M…
#>  6            6 silhouette_avg standard     0.425     5  0.0350 Preprocessor1_M…
#>  7            7 silhouette_avg standard     0.458     5  0.0207 Preprocessor1_M…
#>  8            8 silhouette_avg standard     0.465     5  0.0219 Preprocessor1_M…
#>  9            9 silhouette_avg standard     0.396     5  0.0240 Preprocessor1_M…
#> 10           10 silhouette_avg standard     0.400     5  0.0228 Preprocessor1_M…

^{Created on 2023-03-20 with reprex v2.0.2}

[Ginseng7]

Thanks! you were right!

problem was version of tidyclust.

After tidyclust updated, It worked!
sorry for stupidity.

Thanks for your reply and great work!