Graph Examples

Examples from CausalFusion.

backdoor = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y Adjusted: N/A

frontdoor = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y Adjusted: N/A

frontdoor_backdoor = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y Adjusted: N/A

instrumental_variable = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y

napkin = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y

generate_napkin_data(num_samples: int, treatments: dict[Variable, float] | None = None, *, seed: int | None = None) → DataFrame[source]

Generate testing data for the napkin graph.

Parameters:

num_samples – The number of samples to generate. Try 1000.
treatments – An optional dictionary of the values to fix each variable to. The keys in this dictionary must correspond to variables in the napkin graph as defined in y0.examples.napkin (i.e., with y0.dsl.Z1, y0.dsl.Z2, y0.dsl.X, and y0.dsl.Y).
seed – An optional random seed for reproducibility purposes

Returns:

A pandas Dataframe with columns corresponding to the four variable names in the Napkin graph (i.e., Z1, Z2, X, and Y)

Generate _observational_ data with the following:

>>> from y0.examples.napkin_example
>>> napkin_example.generate_data(1000)

Generate interventional data on \(X=1\) with the following:

>>> from y0.dsl import X
>>> napkin_example.generate_data(1000, treatments={X: 1})

Multiple treatments can be specified:

>>> from y0.dsl import X, Z1
>>> napkin_example.generate_data(1000, treatments={X: 1, Z1: 0})

m_graph = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y Reference:

identifiability_1 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y

identifiability_2 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): Treatment: X Outcome: Y

identifiability_3 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability 3 example Treatment: X Outcome: Y Reference: J. Pearl. 2009. “Causality: Models, Reasoning and Inference. 2nd ed.” Cambridge University Press, p. 92.

identifiability_4 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability 4 example Treatment: X Outcome: Y Reference: J. Pearl. 2009. “Causality: Models, Reasoning and Inference. 2nd ed.” Cambridge University Press, p. 92.

identifiability_5 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability 5 example Treatment: X1, X2 Outcome: Y Reference: J. Pearl. 2009. “Causality: Models, Reasoning and Inference. 2nd ed.” Cambridge University Press, p. 119.

identifiability_6 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability 6 example Treatment: X1, X2 Outcome: Y Reference: J. Pearl. 2009. “Causality: Models, Reasoning and Inference. 2nd ed.” Cambridge University Press, p. 125.

identifiability_7 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability 7 example Treatment: X Outcome: Y Reference: J. Tian. 2002. “Studies in Causal Reasoning and Learning.” p. 90.

verma_1 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Verma 1 example Treatment: V3 Outcome: V4 Reference: T. Verma and J. Pearl. 1990. “Equivalence and Synthesis of Causal Models.” In P. Bonissone et al., eds., Proceedings of the 6th Conference on Uncertainty in Artificial Intelligence. Cambridge, MA: AUAI Press, p. 257.

verma_2 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Verma 2 example Treatment: V1 Outcome: V5 Reference: J. Tian. 2002. “Studies in Causal Reasoning and Learning.” p. 70.

verma_3 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Verma 3 example Treatment: V1 Outcome: V5 Reference: J. Tian. 2002. “Studies in Causal Reasoning and Learning.” p. 59.

verma_4 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Verma 4 example Treatment: V1 Outcome: V5 Reference: E. Bareinboim modification of Verma 2.

verma_5 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Verma 5 example Treatment: V1 Outcome: V5 Reference: E. Bareinboim modification of Verma 2.

z_identifiability_1 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The z-Identifiability 1 example Treatment: X Outcome: Y Z*: Z Reference: E. Bareinboim and J. Pearl. 2012. “Causal Inference by Surrogate Experiments: z-Identifiability.” In Nando de Freitas and K. Murphy., eds., Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence. Corvallis, OR: AUAI Press, p. 114.

z_identifiability_2 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The z-Identifiability 2 example Treatment: X Outcome: Y Z*: Z Reference: E. Bareinboim and J. Pearl. 2012. “Causal Inference by Surrogate Experiments: z-Identifiability.” In Nando de Freitas and K. Murphy., eds., Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence. Corvallis, OR: AUAI Press, p. 114.

z_identifiability_3 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The z-Identifiability 3 example Treatment: X Outcome: Y Z*: Z Reference: E. Bareinboim and J. Pearl. 2012. “Causal Inference by Surrogate Experiments: z-Identifiability.” In Nando de Freitas and K. Murphy., eds., Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence. Corvallis, OR: AUAI Press, p. 114.

identifiability_linear_1 = NxMixedGraph(directed=<networkx.classes.digraph.DiGraph object>, undirected=<networkx.classes.graph.Graph object>): The Identifiability (Linear) 1 example Treatment: X Outcome: Y Reference: J. Pearl. 2009. “Causality: Models, Reasoning and Inference. 2nd ed.” Cambridge University Press, p. 153.