create_abc_settings_object: Creates ensemble-specific parameters for ABC analysis

Description Usage Arguments Value

Description

The EasyABC model wrapper can only take one parameter input: the parameter values. This is problematic as to generate a prediction for those values, we must provide the names of the simulation parameters and measures, the built ensemble, and whether or not the parameter set and responses have been normalised. To get around that problem, this method creates an object in the working directory that contains these values, and the ensemble abc wrapper provided in spartan can then read these in. Thus, this method MUST be run before using the EasyABC package with the ensemble

Usage

1
2
3
create_abc_settings_object(parameters, measures, built_ensemble,
  normalise_values = FALSE, normalise_result = FALSE,
  file_out = TRUE)

Arguments

parameters

Array containing the names of the parameters for which posterior distributions will be generated

measures

Names of the simulation output responses which the ensemble predicts

built_ensemble

An ensemble object created by spartan

normalise_values

Whether the data provided by the EasyABC algorithm should be normalised (as the ensemble must take data between 0 and 1). More than likley this is TRUE, to ensure the posterior distributions are presented in their correct scale

normalise_result

Whether the results produced in running abc generated parameter sets using the ensemble should be rescaled.

file_out

Whether the settings file should be output to file (TRUE) or as an R object (FALSE)

Value

Object containing attributes needed for abc analysis


kalden/spartan documentation built on May 31, 2019, 11:52 p.m.