Micro Free Course — 100% Free Learning
All lessons in this module are free to learn. Sign in with Google to save your progress.
M1 — Regularized Regression + Validation Discipline
Ridge/lasso case, GroupKFold leakage lab, validation quiz, CV notebook + leakage-audit submissions.
Module Content
Regularized-Regression Case (Housing/Price)
Fit ridge/lasso, explain shrinkage effect. Housing train/holdout; leakage trap inside.
Leakage Control + GroupKFold Lab
Outcome: implement GroupKFold; name 3 leakage sources. Curated video (7:19, verified 2026-09-24 via yt-dlp): Stop User Leakage: Group Splits (Professor Py: AI Foundations) Watch: https://www.youtube.com/watch?v=MnuSEB1Yihg Notes: User-level leakage: same entity must never span train/test. Group splits (sklearn GroupKFold / GroupShuffleSplit) as the fix. Lab: implement GroupKFold; name 3 leakage sources (entity overlap, temporal leak, target-derived features). Key moments: No chapter markers on source; key beats: leakage demo, group-split fix, GroupKFold usage.
Quiz: Validation + Leakage
10 scenario MCQs (CV/leakage/group/time splits). Pass >= 80.
CV-Lab Notebook
Grouped/time CV comparison + choice. Pipeline skeleton + 100-pt rubric inside.
Leakage-Audit Writeup
Audit a leaky pipeline, fix + re-score. Deterministic holdout-delta check.