abc_cv: Cross-validation for ABC model

In eam: Evidence Accumulation Models

Cross-validation for ABC model

Description

Wrapper around cv4abc to perform cross-validation of ABC results. This function provides a consistent interface within the eam package and encapsulates the dependency on the abc package.

Usage

abc_cv(abc_input, abc_result, nval, tols, ...)

Arguments

abc_input	A list with components param and sumstat (typically produced by build_abc_input)
abc_result	Fitted ABC model from abc_abc. Parameters like method, transf, etc. are extracted from this object.
nval	Number of cross-validation folds
tols	Tolerance levels to test during cross-validation
...	Additional arguments passed to cv4abc

Details

This is a thin wrapper around the abc::cv4abc() function. When abc_result is provided, cv4abc extracts the method, transf, and other settings from the fitted ABC object. Users should refer to the abc package documentation for detailed parameter descriptions and options.

Value

A cross-validation object from cv4abc

Examples

# Load example simulation output and observed data
rdm_minimal_example <- system.file("extdata", "rdm_minimal", package = "eam")
sim_output <- load_simulation_output(file.path(rdm_minimal_example, "simulation"))
obs_df <- read.csv(file.path(rdm_minimal_example, "observation", "observation_data.csv"))

# Define a summary-statistics pipeline
summary_pipe <- summarise_by(
  .by = c("condition_idx"),
  rt_mean = mean(rt)
)

# Summarise simulation output and observed data
sim_summary <- map_by_condition(
  sim_output,
  .progress = FALSE,
  .parallel = FALSE,
  function(cond_df) {
    summary_pipe(cond_df)
  }
)
obs_summary <- summary_pipe(obs_df)

# Build ABC input and fit an ABC model
abc_input <- build_abc_input(
  simulation_output = sim_output,
  simulation_summary = sim_summary,
  target_summary = obs_summary,
  param = c("V_beta_1", "V_beta_group")
)
abc_model <- abc_abc(
  abc_input = abc_input,
  tol = 0.5,
  method = "rejection"
)

# Run cross-validation for the fitted ABC model
abc_cv_result <- abc_cv(
  abc_input = abc_input,
  abc_result = abc_model,
  nval = 10,
  tols = c(0.1, 0.5)
)

eam documentation built on March 29, 2026, 5:07 p.m.