AncestralGraph

Overview

class graphical_models.classes.mags.ancestral_graph.AncestralGraph(nodes: Set = frozenset({}), directed: Set = frozenset({}), bidirected: Set = frozenset({}), undirected: Set = frozenset({}))

Base class for ancestral graphs, used to represent causal models with latent variables.

Copying

AncestralGraph.copy()	Return a copy of this ancestral graph.
AncestralGraph.induced_subgraph(nodes)	Return the induced subgraph over only nodes

Information about nodes

AncestralGraph.parents_of(nodes)	Return the parents of the node or set of nodes nodes.
AncestralGraph.children_of(i)	Return the children of the node or set of nodes i.
AncestralGraph.spouses_of(nodes)	Return the spouses of the node or set of nodes nodes.
AncestralGraph.neighbors_of(nodes)	Return the neighbors of the node or set of nodes nodes.
AncestralGraph.descendants_of(nodes[, ...])	Return the descendants of the node or set of nodes nodes.
AncestralGraph.ancestors_of(nodes[, ...])	Return the ancestors of the node or set of nodes nodes.
AncestralGraph.district_of(node[, node_subset])	Return the district of a node, i.e., the set of nodes reachable by bidirected edges.
AncestralGraph.markov_blanket_of(node[, flat])	Return the Markov blanket of a node with respect to the whole graph.

Graph modification

AncestralGraph.add_node(node)	Add a node to the ancestral graph.
AncestralGraph.remove_node(node[, ignore_error])	Remove node.
AncestralGraph.add_directed(i, j)	Add a directed edge from node i to node j.
AncestralGraph.remove_directed(i, j[, ...])	Remove the directed edge from i to j.
AncestralGraph.add_bidirected(i, j)	Add a bidirected edge between nodes i and j.
AncestralGraph.remove_bidirected(i, j[, ...])	Remove the bidirected edge between i and j.
AncestralGraph.add_undirected(i, j)	Add an undirected edge between nodes i and j.
AncestralGraph.remove_undirected(i, j[, ...])	Remove the undirected edge between i and j.
AncestralGraph.add_nodes_from(nodes)	Add nodes to the ancestral graph.
AncestralGraph.remove_edge(i, j[, ignore_error])	Remove the edge between i and j, regardless of edge type.
AncestralGraph.remove_edges(edges[, ...])	Remove all edges in edges from the graph, regardless of edge type.

Graph properties

AncestralGraph.legitimate_mark_changes([...])	Return directed edges that can be changed to bidirected edges, and bidirected edges that can be changed to directed edges.
AncestralGraph.discriminating_triples([verbose])	Return the discriminating triples of the graph, which are triples of nodes that determine the discriminating paths.
AncestralGraph.discriminating_paths([verbose])	TODO
AncestralGraph.is_maximal([new, verbose])	TODO
AncestralGraph.c_components()	Return the c-components of this graph.
AncestralGraph.colliders()	TODO
AncestralGraph.vstructures()	TODO
AncestralGraph.has_directed(i, j)	Check if this graph has the directed edge i->``j``.
AncestralGraph.has_bidirected(i, j)	Check if this graph has a bidirected edge between i and j.
AncestralGraph.has_undirected(i, j)	Check if this graph has an undirected edge between i and j.
AncestralGraph.has_any_edge(i, j)	Check if i and j are adjacent in this graph.

Ordering

AncestralGraph.topological_sort()

Return a linear order that is consistent with the partial order implied by ancestral relations of this graph.

Comparison to other AncestralGraphs

AncestralGraph.shd_skeleton(other)	Compute the structure Hamming distance between the skeleton of this graph and the skeleton of another graph.
AncestralGraph.markov_equivalent(other)	Check if this graph is Markov equivalent to the graph other.
AncestralGraph.is_imap(other[, certify])	Check if this graph is an IMAP of the graph other, i.e., all m-separation statements in this graph are also m-separation statements in other.
AncestralGraph.is_minimal_imap(other[, ...])	TODO

Separation Statements

AncestralGraph.msep(A, B[, C])	Check whether A and B are m-separated given C, using the Bayes ball algorithm.
AncestralGraph.msep_from_given(A[, C])	Find all nodes m-separated from A given C.

Conversion to/from other formats

AncestralGraph.to_amat()	Convert the graph into an adjacency matrix.
AncestralGraph.from_amat(amat)	Create a graph from an adjacency matrix.