geoutils.PropagationSummary

class geoutils.PropagationSummary(*, estimate, mean, std, error_structure, method, selection, output=dict(), samples=None, quantiles=None, n_samples=None, n_success=None, n_valid=None, failures=dict(), resultant_length=None, metadata=dict())

Propagation summary that stores propagated uncertainty descriptors, whether obtained analytically or numerically.

The PropagationSummary class is returned by propagate(). estimate is the result from the original values; mean and std describe the result after adding source errors. With zero-mean errors, the expected result of a weighted sum equals estimate; a nonlinear calculation can shift it. samples optionally stores random draws used in numerical propagation. Use interval() for uncertainty bounds or quantile() for a percentile.

Parameters:

• estimate (Any) – Output calculated before adding source errors.

• mean (Any) – Expected output after adding source errors, estimated from draws for numerical propagation.

• std (Any) – Standard deviation of the results after including source uncertainty.

• error_structure (ErrorStructure) – Error model used for the source observations.

• method (Literal['analytical', 'numerical']) – “analytical” for exact weighted sums, or “numerical” for repeated random draws (Monte Carlo).

• selection (DataFrame) – Labels and original estimates for selected samples or quantiles.

• samples (DataFrame | None) – Optional output values calculated from random draws of the source errors.

• n_samples (int | None) – Number of random draws requested.

• n_valid (Any) – Number of finite draws at each output.

• metadata (Mapping[str, object]) – Details about the calculation and any approximation it used.

__init__(*, estimate, mean, std, error_structure, method, selection, output=<factory>, samples=None, quantiles=None, n_samples=None, n_success=None, n_valid=None, failures=<factory>, resultant_length=None, metadata=<factory>)

Methods

__init__(*, estimate, mean, std, ...[, ...])
interval([coverage])	Return uncertainty bounds with equal probability below and above them.
marginal(at)	Return the uncertainty distribution for one output.
quantile(probability)	Return a quantile from saved draws or an explicitly known normal marginal.

Attributes

bias	Return the difference between the mean and the result from the original source values.
n_samples
n_success
n_valid
quantiles
resultant_length
samples
variance	Return the variance of each output.
estimate
mean
std
error_structure
method
selection
output
failures
metadata