tangram.utils.eval_metric

tangram.utils.eval_metric(df_all_genes, test_genes=None)

Compute metrics on given test_genes set for evaluation

Parameters

• df_all_genes (Pandas dataframe) – returned by compare_spatial_geneexp(adata_ge, adata_sp);

• test_genes (list) – list of test genes, if not given, test_genes will be set to genes where ‘is_training’ field is False

Returns

dict with values of each evaluation metric (“avg_test_score”, “avg_train_score”, “auc_score”), tuple of auc fitted coordinates and raw coordinates(test_score vs. sparsity_sp coordinates)