R/simple_ensemble.R

In tidysdm: Species Distribution Models with Tidymodels

#' Simple ensemble
#'
#' A simple ensemble is a collection of workflows for which predictions will be
#' combined in a simple way (e.g. by taking either the mean or median). Usually
#' these workflows will consists each of the best version of a given model
#' algorithm following tuning. The workflows are fitted to the full training
#' dataset before making predictions.
#'
#' @param ... not used, this function just creates an empty `simple_ensemble`
#'   object. Members are added with `add_best_candidates()`
#' @returns an empty `simple_ensemble`. This is a tibble with columns:
#' * `wflow_id`: the name of the workflows for which the best model was
#'   chosen
#' * `workflow`: the trained workflow objects
#' * `metrics`: metrics based on the crossvalidation resampling used
#'   to tune the models
#' @export

simple_ensemble <- function(...) {
  parsnip::check_empty_ellipse(...)
  # a tibble with columns: name, workflow
  base_ensemble <- tibble::tibble(
    wflow_id = character(),
    workflow = list(),
    # tibble of metrics from the CV on the training dataset (coming from when
    # the workflow was originally fit, potentially as part of a workflow_set)
    metrics = list()
  )

  base_ensemble <- structure(base_ensemble, class = c(
    "simple_ensemble",
    class(base_ensemble)
  ))
}

#' @export
print.simple_ensemble <- function(x, ...) {
  rlang::inform("A simple_ensemble of models")
  if (nrow(x) > 0) {
    rlang::inform(c("\nMembers:", x$wflow_id))
    rlang::inform(c("\nAvailable metrics:", attr(x, "metrics")))
    rlang::inform(c("\nMetric used to tune workflows:", attr(x, "best_metric")))
  } else {
    rlang::inform("\nThis object is empty; add models with `add_members()`")
  }
}

#' @export
summary.simple_ensemble <- function(object, ...) {
  print(object)
}