Translating from {tidymodels} and scikit-learn: Lessons from a 'bilingual' course
summary
A course translates one curriculum into both tidymodels (R) and scikit-learn (Python) and finds concrete mismatches: a single-column extract is 1D in R but a 1-column DataFrame in Python; tidymodels recipes group predictor choice with transforms while scikit-learn pipelines apply preprocessing first; SVM’s penalty is parameterized oppositely (C vs cost inversion); kmeans defaults (Hartigan‑Wong vs kmeans++) produce different clusters. Recommendation: document defaults, state parameter meaning in plain language, avoid wrapper shortcuts for core learning goals.
AI-written summary · about
chapters (5)
- 0:00Start
- 10:18"FALSE FRIENDS"
- 11:04SUPPORT VECTOR MACHINES
- 13:54SLANG TERMS
- 16:38NONVERBAL COMMUNICATION
related talks
01h 24mInpsect and try to interpret your scikit-learn machine-learning models
27mSoftware Library APIs: Lessons Learned from scikit-learn
01h 17mMachine Learning with Scikit-Learn
35mPandas, Pipelines, and Custom Transformers
30mKick-Start your Understanding of Machine Learning with Python
30mPatsy: The Lingua Franca to and from R
32mMachine Learning Infrastructure at Stripe: Bridging from Python JVM
27mUsing Machine Learning to solve a classification problem with scikit-learn