Quickstart
This guide provides a minimal, complete example of how to use citk
for a conditional independence test within the causal-learn
framework.
Example: Running PC with a CITK Test
The following example demonstrates how to:
Generate synthetic data.
Run the PC algorithm.
Use a conditional independence test from
citk
.
import numpy as np
from causallearn.search.ConstraintBased.PC import pc
import citk.tests
# 1. Generate some data
# Here, X2 is a function of X0 and X1, creating a dependency.
np.random.seed(42)
data = np.random.randn(200, 3)
data[:, 2] = 0.5 * data[:, 0] + 0.5 * data[:, 1] + 0.1 * np.random.randn(200)
# 2. Run the PC algorithm using a citk test
# You can swap 'spearman' with any other available test, like
# "fisherz", "gsq", "rf", "dml", etc.
cg = pc(data, alpha=0.05, indep_test='spearman')
# 3. View the learned graph edges
# The output should reflect the dependencies in the data.
print("Learned Graph Edges:")
print(cg.G.get_edges())