Module causal_inference 

Causal inference and structural causal models.

Provides:

• CausalGraph — directed acyclic graph, topological sort, d-separation
• StructuralCausalModel — linear SCM with noise terms, do-calculus interventions
• BackdoorCriterion — backdoor criterion check and adjustment
• FrontdoorCriterion — frontdoor adjustment formula
• PropensityScoreMatching — propensity score estimation, ATT/ATE
• InstrumentalVariables — IV estimation, two-stage least squares (2SLS)
• CausalDiscovery — PC algorithm skeleton, orientation rules
• CounterfactualQuery — E[Y|do(X=x), Z=z] style queries

Structs

BackdoorCriterion

Checks whether a set of variables satisfies the backdoor criterion for identifying the causal effect of treatment on outcome.

CausalDiscovery

Causal discovery via the PC algorithm.

CausalGraph

A directed acyclic graph (DAG) representing causal relationships between variables.

CounterfactualQuery

Compute counterfactual queries of the form E[Y | do(X=x), Z=z].

FrontdoorCriterion

Implements the frontdoor adjustment formula for causal effect identification.

InstrumentalVariables

Instrumental variables (IV) estimation and Two-Stage Least Squares (2SLS).

PropensityScoreMatching

Propensity score matching for observational causal inference.

StructuralCausalModel

A linear structural causal model (SCM).

Functions

fisher_z_test

Fisher Z-test for conditional independence.

partial_correlation

Compute partial correlation of variables i and j conditioning on cond.

pearson_correlation

Pearson correlation between variables i and j in a data matrix.

sample_covariance

Compute the sample covariance matrix from a data matrix.