Skip to main content

Causal Effects

@validate_call
def causal_effects(actions: Union[str, dict[str, tuple[float, float]]],
                   fixed: dict[str, float] = None,
                   interval: float = 0.90,
                   observation_noise=False) -> pd.DataFrame

Get the causal effects of actions on the model.

Arguments:

  • actions Union[str, dict[str, tuple[np.ndarray, np.ndarray]]] - A dictionary representing the actions.
  • fixed dict - A dictionary representing the fixed nodes.
  • interval float - The interval at which to simulate the action.
  • observation_noise bool - Whether to include observation noise.

Returns:

  • pd.DataFrame - A dataframe representing the causal effects of the actions.

Example:

model.causal_effects( ... {"x": [0, 1]} ... )