With recipes, you can use dplyr-like pipeable sequences of feature engineering steps to get your data ready for modeling. For example, to create a recipe containing an outcome plus two numeric predictors and then center and scale (“normalize”) the predictors:
library(recipes) data(ad_data, package = "modeldata") ad_rec <- recipe(Class ~ tau + VEGF, data = ad_data) %>% step_normalize(all_numeric_predictors()) ad_rec #> Recipe #> #> Inputs: #> #> role #variables #> outcome 1 #> predictor 2 #> #> Operations: #> #> Centering and scaling for all_numeric_predictors()
You may consider recipes as an alternative method for creating and preprocessing design matrices (also known as model matrices) that can be used for modeling or visualization. While R already has long-standing methods for creating such matrices (e.g. formulas and
model.matrix), there are some limitations to what the existing infrastructure can do.
There are several ways to install recipes:
For questions and discussions about tidymodels packages, modeling, and machine learning, please post on RStudio Community.
If you think you have encountered a bug, please submit an issue.
Either way, learn how to create and share a reprex (a minimal, reproducible example), to clearly communicate about your code.