PosteriorEstimator: PosteriorEstimator

In NeuralEstimators: Likelihood-Free Parameter Estimation using Neural Networks

PosteriorEstimator

Description

Construct a neural posterior estimator. The neural-network architecture is still defined in Julia (e.g. via juliaEval()); this function wraps that architecture in a PosteriorEstimator.

Typical constructors, matching the Julia methods:

• PosteriorEstimator(num_parameters, summary_network, num_summaries = ...): builds the approximate distribution internally (default q = "NormalisingFlow").

• PosteriorEstimator(summary_network, q): an explicit approximate distribution q.

The argument q may be a Julia type or a string naming one (e.g. "Gaussian", "GaussianMixture", "NormalisingFlow").

Usage

PosteriorEstimator(...)

Arguments

...	arguments passed to the Julia version of PosteriorEstimator()

Value

a PosteriorEstimator

See Also

PointEstimator(), RatioEstimator(), train(), infer()

NeuralEstimators documentation built on Sept. 10, 2026, 1:12 a.m.