workflowsets

The workflowsets package allows users to create and fit multiple machine learning models simultaneously by organizing them into workflow sets. It generates combinations of preprocessors (like recipes and formulas) and model specifications, then provides functions to tune or resample all combinations efficiently.

This package solves the problem of testing different preprocessing approaches and model types on the same dataset without manually creating each combination. It integrates with the tidymodels framework to support systematic model comparison through grid search, resampling, and performance ranking. Users can easily filter out illogical combinations (like pairing PCA with regularized models) and evaluate all remaining workflows with a single function call.