Create a tidymodels workflow for H3-based SDMs

Combines a model specification and a prepared recipe into a single tidymodels workflow. This workflow is suitable for species distribution modeling using H3 hexagonal grids and can be directly fitted or cross-validated.

Usage

h3sdm_workflow(model_spec, recipe)

Arguments

model_spec: A tidymodels model specification (e.g., logistic_reg(), rand_forest(), or boost_tree()), describing the model type and engine to use for fitting. Use set_mode("classification") for presence/absence models and set_mode("regression") for count-based models (species richness, detections, or individuals).
recipe: A tidymodels recipe object, typically created with h3sdm_recipe(), which preprocesses the data and defines predictor/response roles. Use response_col = "count" in h3sdm_recipe() when working with count data.

Value

A workflow object ready to be used for model fitting with fit() or cross-validation.

Details

The function creates a workflow that combines preprocessing and modeling steps. This encapsulation allows consistent model training and evaluation with tidymodels functions like fit() or fit_resamples(), and is particularly useful when applying multiple models in parallel.

Choosing the model mode:

For presence/absence data: use set_mode("classification").
For count data (species richness, detections, individuals): use set_mode("regression").

Examples

if (FALSE) { # \dontrun{
library(parsnip)

# --- Presence/absence model ---
rf_spec_pa <- rand_forest() %>%
  set_engine("ranger") %>%
  set_mode("classification")

rec_pa <- h3sdm_recipe(combined_data)

wf_pa <- h3sdm_workflow(model_spec = rf_spec_pa, recipe = rec_pa)

# --- Count-based model ---
rf_spec_count <- rand_forest() %>%
  set_engine("ranger") %>%
  set_mode("regression")

rec_count <- h3sdm_recipe(combined_data, response_col = "count")

wf_count <- h3sdm_workflow(model_spec = rf_spec_count, recipe = rec_count)
} # }