Calculate cluster entropy for a categorical label — sn_calculate_cluster

Cluster entropy measures how mixed a categorical label is within each cluster. When used with batch labels, higher normalized entropy indicates stronger within-cluster batch mixing.

Usage

sn_calculate_cluster_entropy(x, cluster, label)

Arguments

x: A Seurat object or data frame containing the required columns.
cluster: Metadata/data-frame column containing cluster labels.
label: Metadata/data-frame column containing the label to evaluate within each cluster.

Value

A data frame with one row per cluster, including raw entropy and a normalized entropy score in [0, 1].

Examples

meta <- data.frame(
  cluster = c("0", "0", "1", "1"),
  batch = c("a", "b", "a", "b")
)
sn_calculate_cluster_entropy(meta, cluster = "cluster", label = "batch")
#>   cluster n_cells n_labels dominant_label   entropy normalized_entropy
#> 1       0       2        2              a 0.6931472                  1
#> 2       1       2        2              a 0.6931472                  1