sinfonia.metrics.clustering_metrics

sinfonia.metrics.clustering_metrics(adata, target, pred)

Evaluate clustering performance.

Parameters

• adata – AnnData object of shape n_obs × n_vars. Rows correspond to cells and columns to genes.

• target – Key in adata.obs where ground-truth spatial domain labels are stored.

• pred – Key in adata.obs where clustering assignments are stored.

Returns

• ami – Adjusted mutual information score.

• ari – Adjusted Rand index score.

• homo – Homogeneity score.

• nmi – Normalized mutual information score.