R/data.R

In D-Se/ML: Machine Leanring targets-bookdown pipeline

#' @title Object specification of the ML pipeline
#'
#' @description A collection of objects along the critical path of analysis process
#'   useful for model prototyping and verification.
#'
#' @format A list containing three elements
#' \describe{
#'   \item{splits}{an `rsplit` object of the `rsample` package that indicates splitting criteria and data inclusion}
#'   \item{recipes}{assortment of `recipes` package recipe objects describing preprocessing steps}
#'   \item{models}{a list of `parsnip` model specifications to see hyperparameter tuning criteria}
#' }
#' @source \url{https://github.com/D-Se/ML}
"analysis"
# analysis = list(
#   splits = targets::tar_read(init),
#   recipes = targets::tar_read(recipes),
#   models = targets::tar_read(models)
# )




#' @title temporary airquality data for bookdown
#' @format see `airquality`
#' @source New York State Department of Conservation (ozone data) and the National Weather Service (meteorological data).
data <- NULL


#' @title Best performing ML model in group project
#'
#' @description A quick insight into the best performing model of the bunch
#'
#' @format A list containing three elements
#' \describe{
#'   \item{predictions}{a tibble with model performance predictions and observed events.}
#'   \item{workflow}{a `workflow` object that specifies exact hyperparameter settings of the best model}
#' }
#' @source \url{https://github.com/D-Se/ML}
"results"

# results = list(
#   predictions = targets::tar_read(predictions)[,c(3, 2, 4)] |>
#     setNames(c("sample_row", "model_prediction", "actual_observation")),
#   workflow = targets::tar_read(winners_post_test)$.workflow[[1]]
# )