Run full train_model workflow using multiple modeling approaches

Usage

train_models(
  prep_output,
  models,
  y_var,
  pred_vars,
  id_vars = c("orgUnit", "date"),
  results_dir = NULL,
  tune = NULL,
  model_configs = NULL,
  create_report = FALSE,
  report_configs = NULL
)

Arguments

prep_output: output of prep_data
models: vector of models to fit. Options: naive, arimax, glm_nb, ranger, inla
y_var: character. name of variable to predict
pred_vars: character vector of predictor variables. Not all will be used in all models
results_dir: directory to save model results and quarto document in. If NULL, temporary directory will be used
tune: string vector containing names of models to tune. Options = ranger
model_configs: Optional list of configurations for each model. Name of element in list should match model name. See ensemble_forecast for more info.
create_report: Whether to create the HTML report. Default = FALSE
report_configs: Optional configurations for quarto_report provided as a named list. Options are: html_filename, doc_title, lang

Value

saves model outputs to results_dir. Creates a quarto doc of model outputs saved in results_dir