idmodelr (version 0.4.0)

generate_parameter_permutations: A Function to Generate Parameter Permutations

Description

A function to generate parameter permutations from a generic sampling function (or if not given from the input parameters). This function can be used to rapidly generate new parameter combinations given parameters to be varied, and scenarios to be investigated.

Usage

generate_parameter_permutations(
  variable_params = NULL,
  fixed_params = NULL,
  sample_params = NULL,
  excluded_params = NULL,
  scenarios = NULL,
  sampling_function = NULL,
  parameter_samples = 1,
  repeat_sample = TRUE,
  rerun = FALSE,
  ...
)

Value

A dataframe containing sampled parameter permutations

Arguments

variable_params

A dataframe containing any parameter variations to be investigated. If these parameters would normally be sampled then they should be added to the excluded_params argument.

fixed_params

A named vector of parameters that are not sampled by the sampling function. If these parameters would usually be sampled then they should be added to the excluded_params argument.

sample_params

A named vector of parameters to be sampled. If a sampling function is not supplied these parameters will be used in the final permutation dataframe.

excluded_params

A character vector indicating which parameters should have there sampled values kept.

scenarios

A dataframe of possible scenarios to investigate. It must contain a scenario variable to identify each separate scenarios. If parameters are included here that would normally be sampled then they should be added to the excluded_params argument

sampling_function

A sampling function, this should be designed such that it's input is a matrix with each parameter having a named row. It should return it's output in the same format. If not supplied defaults to passing through parameters, this may not be the required behaviour.

parameter_samples

The number of parameter samples to take, defaults to one.

repeat_sample

A logical (defaults to TRUE) which indicates if each scenario should independently sample from the sampling function. If set to FALSE then each scenario will share the same sampled parameter set.

rerun

A logical indicating if the function should be rerun or saved results should be loaded. Defaults to FALSE.

...

Additional arguments to be passed to the sampling_function.

Examples

scenarios <- data.frame(scenario = c("test_1", "test_2"), scenario_param = c(0, 1))
variable_params <-  data.frame(variable = c(0, 0.5, 1))
fixed_params <- c(fixed_1 = 2, fixed_2 = c(1, 3, 4))
sample_params <- c(sample_1 = 2, sample_2 = c(2, 1))

generate_parameter_permutations(variable_params, fixed_params, sample_params,
                                excluded_params = c("variable"), scenarios,
                                parameter_samples = 1)