Modelling, Tidyverse Way
summary
Tidy modeling shows a concrete workflow: load resample to do an 80/20 stratified split on the 150‑row Auto MPG set, use recipes (step_ functions) to collapse rare factor levels, impute and one‑hot encode, then parsnip to declare a model type separately from the engine and fit it, getting a tibble of predictions. Version noted as 0.02–0.03; everything runs first time, and yardstick* returns tidy metric tables for straightforward comparison.
AI-written summary · about
chapters (9)
- 0:00Introduction to Mango
- 2:11The challenges of R
- 4:42The need for tidiness
- 7:18Modeling in traditional R
- 8:44Introducing tidymodels
- 10:55Data splitting demo
- 14:12Pre-processing with recipes
- 18:03Modeling with parsnip
- 20:44Prediction and metrics
related talks
57mGood Practices for Applied Machine Learning
18mtidymodels: Adventures in Rewriting a Modeling Pipeline
23mTotal Tidy Tuning Techniques- 57mGood Practices for Applied Machine Learning
21mSolving the model representation problem with broom
02h 16mTech Tutorial: Create Twitter and APIs
15mThe tidysynthesis R package
51mProgramming in the Tidyverse