Skip to content
voiage

voiage.backends.enhanced_jax_backend.EnhancedJaxBackend

Enhanced JAX backend with advanced optimization features.

Methods

Section titled “Methods”

evppi_advanced

Section titled “evppi_advanced”
evppi_advanced([positional or keyword] self: None = None, [positional or keyword] net_benefit_array: np.ndarray = None, [positional or keyword] parameter_samples: np.ndarray | Mapping[str, np.ndarray] = None, [positional or keyword] parameters_of_interest: Sequence[str] = None, [positional or keyword] method: str = 'polynomial', [positional or keyword] degree: int = 2, [positional or keyword] cv_folds: int = 5, [positional or keyword] regularization: float = 1e-06) -> float

Advanced EVPPI calculation with enhanced regression models.

Args: net_benefit_array: Net benefit data parameter_samples: Parameter samples parameters_of_interest: Parameters to analyze method: Regression method (“polynomial”, “ridge”, “lasso”) degree: Polynomial degree for polynomial regression cv_folds: Number of cross-validation folds regularization: Regularization parameter

Parameters:

  • self
  • net_benefit_array np.ndarray
  • parameter_samples np.ndarray | Mapping[str, np.ndarray]
  • parameters_of_interest Sequence[str]
  • method str (default: 'polynomial')
  • degree int (default: 2)
  • cv_folds int (default: 5)
  • regularization float (default: 1e-06)

Returns: float

batch_evppi

Section titled “batch_evppi”
batch_evppi([positional or keyword] self: None = None, [positional or keyword] net_benefit_arrays: Sequence[np.ndarray] = None, [positional or keyword] parameter_samples: np.ndarray | Mapping[str, np.ndarray] = None, [positional or keyword] parameters_of_interest: Sequence[str] = None) -> np.ndarray

Batch EVPPI calculation for multiple net benefit arrays.

Parameters:

  • self
  • net_benefit_arrays Sequence[np.ndarray]
  • parameter_samples np.ndarray | Mapping[str, np.ndarray]
  • parameters_of_interest Sequence[str]

Returns: np.ndarray

parallel_monte_carlo

Section titled “parallel_monte_carlo”
parallel_monte_carlo([positional or keyword] self: None = None, [positional or keyword] net_benefit_array: np.ndarray = None, [positional or keyword] n_simulations: int = 1000, [positional or keyword] chunk_size: int = 100) -> dict[str, object]

Parallel Monte Carlo sampling for variance reduction.

Parameters:

  • self
  • net_benefit_array np.ndarray
  • n_simulations int (default: 1000)
  • chunk_size int (default: 100)

Returns: dict[str, object]