pystra.results.SensitivityResult

Use it: User guide · Worked example · Theory

class SensitivityResult(*, method, status, message, n_limit_state_evaluations, variable_names, options=None, failure_probability, beta, form, approach, marginal, correlation, delta, diagnostics=dict())

Bases: _ProbabilityResult

Record of a sensitivity analysis of the FORM reliability index.

failure_probability and beta are those of form, the FORM analysis at the model’s parameters. The sensitivities are derivatives of its design index.

form

FORM at the model’s parameters.

Type:

FORMResult

approach

Forward finite differences, or Bourinet’s (2017) closed form.

Type:

{“numerical”, “closed_form”}

marginal

Derivative of the index with respect to each distribution parameter, by variable and parameter name.

Type:

Mapping[str, Mapping[str, float]]

correlation

Closed form only: derivatives with respect to the correlation coefficients, symmetric with a zero diagonal.

Type:

ndarray or None

delta

Numerical only: the relative perturbation.

Type:

float or None

diagnostics

On nonconvergence, failed_form retains the failed inner FORM record and phase identifies the baseline or perturbation. A perturbation also records its variable, parameter and actual step.

Type:

Mapping

Methods

summary	Return a plain-text report of the principal results.
to_dataframe	Return the marginal sensitivities as a tidy table.

to_dataframe()

Return the marginal sensitivities as a tidy table.

Returns:

Columns Variable, Parameter and ∂β/∂θ.

Return type:

DataFrame

property converged: bool

Whether the method met its termination criteria.

summary()

Return a plain-text report of the principal results.

Return type:

str