estimate {NeuralEstimators}R Documentation

estimate

Description

estimate parameters from observed data using a neural estimator

Usage

estimate(estimator, Z, theta = NULL, use_gpu = TRUE)

Arguments

estimator

a neural estimator

Z

data; format should be amenable to the architecture of estimator

theta

parameter vectors (only for neural estimators that take both the data and parameters as input, e.g., neural ratio estimators)

use_gpu

a boolean indicating whether to use the GPU if it is available (default true)

Value

a matrix of parameter estimates (i.e., estimator applied to Z)

Examples

## Not run: 
library("NeuralEstimators")
library("JuliaConnectoR")

## Observed data: 100 replicates of a univariate random variable
Z = matrix(rnorm(100), nrow = 1)

## Construct an (un-trained) point estimator
estimator <- initialise_estimator(1, architecture = "MLP")

## Apply the estimator
estimate(estimator, Z)
## End(Not run)

[Package NeuralEstimators version 0.1.0 Index]