nichecompass.train.eval_metrics
- nichecompass.train.eval_metrics(edge_recon_probs, edge_labels, edge_same_cat_covariates_cat=None, edge_incl=None, omics_pred_dict=None)
Get the evaluation metrics for a (balanced) sample of positive and negative edges and a sample of nodes.
- Parameters:
edge_recon_probs (
Union[Tensor,ndarray]) – Tensor or array containing reconstructed edge probabilities.edge_labels (
Union[Tensor,ndarray]) – Tensor or array containing ground truth labels of edges.edge_incl (
Union[Tensor,ndarray,None] (default:None)) – Boolean tensor or array indicating whether the edge should be included in the evaluation.target_rna_preds – Tensor or array containing the predicted gene expression.
target_rna – Tensor or array containing the ground truth gene expression.
source_rna_preds – Tensor or array containing the predicted gene expression.
source_rna – Tensor or array containing the ground truth gene expression.
chrom_access_preds – Tensor or array containing the predicted chromatin accessibility.
chrom_access – Tensor or array containing the ground truth chromatin accessibility.
- Return type:
- Returns:
eval_dict: Dictionary containing the evaluation metrics ´auroc_score´ (area under the receiver operating characteristic curve), ´auprc score´ (area under the precision-recall curve), ´best_acc_score´ (accuracy under optimal classification threshold) and ´best_f1_score´ (F1 score under optimal classification threshold).