causal_falsify.utils.cond_indep module

causal_falsify.utils.cond_indep.fisherz(x, y, z)[source]

Fisher’s Z Conditional Independence Test.

Parameters:

  • x (np.ndarray) – Input array representing variable X. Must be a 2D array with shape (n_samples, 1).

  • y (np.ndarray) – Input array representing variable Y. Must be a 2D array with shape (n_samples, 1).

  • z (np.ndarray) – Input array representing conditioning variables Z. Must be a 2D array with shape (n_samples, n_covariates).

Returns:

p-value of the test if successful; None if an error occurred.

Return type:

Optional[float]

causal_falsify.utils.cond_indep.kcit_rbf(x, y, z)[source]

Kernel-based Conditional Independence Test (KCIT) with RBF kernels.

Parameters:

  • x (np.ndarray) – Input array representing variable X. Must be a 2D array with shape (n_samples, 1).

  • y (np.ndarray) – Input array representing variable Y. Must be a 2D array with shape (n_samples, 1).

  • z (np.ndarray) – Input array representing conditioning variables Z. Must be a 2D array with shape (n_samples, n_covariates).

Returns:

p-value of the test if successful; None if an error occurred.

Return type:

Optional[float]