synthcity.plugins.core.models.aim module

class AIM(epsilon: float, delta: float, rounds: Optional[Union[int, float]] = None, max_model_size: int = 80, structural_zeros: Dict = {})

Bases: synthcity.plugins.core.models.aim.Mechanism

exponential_mechanism(qualities: Union[Dict, numpy.ndarray, Any], epsilon: float, sensitivity: Union[float, int] = 1.0, base_measure: Optional[Dict] = None) → numpy.ndarray

gaussian_noise(sigma: float, size: Union[int, Tuple]) → numpy.ndarray

Generate iid Gaussian noise of a given scale and size

run(data: synthcity.plugins.core.models.mbi.dataset.Dataset, W: List) → synthcity.plugins.core.models.mbi.dataset.Dataset

worst_approximated(candidates: Dict, answers: Dict, model: synthcity.plugins.core.models.mbi.graphical_model.GraphicalModel, eps: float, sigma: float) → numpy.ndarray

class Mechanism(epsilon: float, delta: float)

Bases: object

exponential_mechanism(qualities: Union[Dict, numpy.ndarray, Any], epsilon: float, sensitivity: Union[float, int] = 1.0, base_measure: Optional[Dict] = None) → numpy.ndarray

gaussian_noise(sigma: float, size: Union[int, Tuple]) → numpy.ndarray

Generate iid Gaussian noise of a given scale and size

run(dataset: synthcity.plugins.core.models.mbi.dataset.Dataset, workload: List[Tuple]) → Any

cdp_delta(rho: Union[float, int], eps: Union[float, int]) → Union[float, int]

cdp_rho(eps: float, delta: float) → float

compile_workload(workload: List[Tuple]) → Dict

default_params() → Dict[str, Any]

Return default parameters to run this program

Returns

a dictionary of default parameter settings for each command line argument

downward_closure(Ws: List[Tuple]) → List

filter_candidates(candidates: Dict, model: synthcity.plugins.core.models.mbi.graphical_model.GraphicalModel, size_limit: Union[float, int]) → Dict

hypothetical_model_size(domain: synthcity.plugins.core.models.mbi.domain.Domain, cliques: List[Union[Tuple, List]]) → float

powerset([1,2,3]) --> (1,) (2,) (3,) (1,2) (1,3) (2,3) (1,2,3)