Transduce k-means cluster centers to shapes

Reconstruct shapes at cluster centers or at positions relative to centers.

Usage

# S3 method for class 'stat_kmeans'
transduce(object, positions)

Arguments

object

A stat_kmeans object

positions

A tibble with cluster specifications. Must contain a cluster column with cluster numbers (1 to k).

Value

A tibble with reconstructed coefficients and shapes

Details

For k-means, transduce simply retrieves the cluster centers from the model and reconstructs the corresponding shapes. Since cluster centers are already in the original predictor space, no inverse transformation is needed.

Examples

if (FALSE) { # \dontrun{
km <- boteft %>% stat_kmeans(k = 3)

# Get shapes at all cluster centers
centers <- transduce(km, tibble(cluster = 1:3))

# Get shape at specific cluster
center1 <- transduce(km, tibble(cluster = 1))
} # }