grassp.tl.calinski_habarasz_score#

calinski_habarasz_score(data, gt_col, use_rep='X_umap', key_added='ch_score', class_balance=False, inplace=True, seed=42)[source]#

Calinski–Harabasz score of cluster compactness vs separation.

Parameters:

data: AnnData with an embedding under .obsm[use_rep].
gt_col: Observation column containing cluster assignments.
use_rep default: 'X_umap': Name of embedding to use (default "X_umap").
key_added default: 'ch_score': Key under which to store the score when inplace is True.
class_balance default: False: If True subsample each cluster to equal size before computing the score (calls class_balance internally).
inplace default: True: Standard behaviour flags.
seed default: 42: Standard behaviour flags.

Return type:

None | float

Returns:

If inplace is True:

If inplace is False:

Score.

grassp.tl.calinski_habarasz_score