Good Practices for Applied Machine Learning
summary
Tidymodels presents a unified R modeling stack (yardstick, tune, recipes, workflows) that encapsulates preprocessing so it applies identically at training and prediction, adds extensible tools (textrecipes, stacks, modeltime) and a deployment layer vetiver that supports R and Python and one-line Plumber APIs. Hyperparameter racing eliminates ~93% of candidates so tuning can run ~3x faster in parallel (9x without). Built workflows prevent data leakage by redoing feature selection inside resamples; website and O’Reilly book provide practical examples.
AI-written summary · about
related talks
57mGood Practices for Applied Machine Learning
18mtidymodels: Adventures in Rewriting a Modeling Pipeline
22mModelling, Tidyverse Way
36mApplied Machine Learning Playbook: From Idea to Production
30mReproducible Machine Learning at Scale
20mBetter Care and Feeding of Machine Learning Models
26mModel Selection with Python: An Introduction to Hyper Parameter Tuning
23mTotal Tidy Tuning Techniques