scButterfly.calculate_cluster.calculate_cluster_index

scButterfly.calculate_cluster.calculate_cluster_index(adata)

Evaluation of cluster index of prediction. Cluster method using Leiden with a default resolution. 4 different metrics provided: ARI - Adjusted Rand Index, NMI - Normalized Mutual Information, AMI - Adjusted Mutual Information, HOM - Homogeneity

Parameters:

adata (Anndata) – Anndata need to cumpute index, there need ground truth labels “cell_type” in adata.obs.

Returns:

• ARI (float) – Adjusted Rand Index.

• AMI (float) – Normalized Mutual Information.

• NMI (float) – Adjusted Mutual Information.

• HOM (float) – Homogeneity.