cascade.metrics

Metrics for evaluating the accuracy of inferred causal structures

Functions

annot_resp	Annotate interventional responsiveness for a predicted causal graph
cmp_true_pred	Compare the true and predicted causal graphs in a long-form data frame
ctfact_delta_pcc	Pearson correlation coefficient of counterfactual delta
ctfact_dir_acc	Directional accuracy of counterfactual predictions
ctfact_mse	Mean squared errors of counterfactual prediction
disc_acc	Accuracy of the predicted causal graph
disc_ap	Average precision of the predicted causal graph
disc_auroc	Area under ROC curve of the predicted causal graph
disc_f1	F1 score of the predicted causal graph
disc_prec	Precision of the predicted causal graph
disc_recall	Recall of the predicted causal graph
disc_resp_acc	Responsiveness accuracy of the predicted causal graph
disc_resp_dist	Responsiveness distance of the predicted causal graph
disc_resp_dist_diff	Responsiveness distance difference of the predicted causal graph
disc_shd	Structural hamming distance between the true and predicted causal graph
disc_sid	Structural interventional distance between the true and predicted causal graph
dsgn_auhrc_exact	Area under the exact hit-rate curve for intervention design, see dsgn_hrc_exact()
dsgn_auhrc_partial	Area under the partial hit-rate curve for intervention design, see dsgn_hrc_partial()
dsgn_hrc_exact	Exact hit-rate curve for intervention design
dsgn_hrc_partial	Partial hit-rate curve for intervention design
optimal_cutoff	Obtain the optimal binary classification cutoff