3.1.6. pywhy_graphs.networkx.bidirected_to_unobserved_confounder#

pywhy_graphs.networkx.bidirected_to_unobserved_confounder(G, directed_edge_name='directed', bidirected_edge_name='bidirected', uc_label='Unobserved Confounders')[source]#

Convert all bidirected edges to unobserved confounders.

Parameters:

G : MixedEdgeGraph

A causal graph with bidirected edges.

directed_edge_name : str

The name of the graph representing directed edges in G.

bidirected_edge_name : str

The name of the graph representing bidirected edges in G.

uc_label : str

The label of the unobserved variables that are added in.

Returns:

G_copy : DiGraph

A networkx DiGraph that is a fully specified DAG with unobserved variables added in place of bidirected edges.

Notes

This converts bidirected to unobserved confounding variables, that are unobserved nodes that have a directed edge pointing to the two variables that were connected with a bidirected edge.