Skip to learning content

DATA SCIENCE PYTHON PLAYGROUND

Machine Learning · Learn / Refresh

← Choose a deckQUESTIONS · MODELS · EVIDENCE
DECK 1

ML Foundations

Questions and Python basics, then complete regression and classification workflows.

01 / Questions and tables

02 / Learning from examples

03 / Honest evaluation

F06Core · teaching
Seen is not unseen: All observations, Seen for learning, Unseen for testAll observationsSeen for learningUnseen for test

Seen is not unseen

Explain why training performance is insufficient.

12–18 min · 3 practices
F07Core · teaching
Making a reproducible split: Aligned X and y, Seeded train rows, Reserved test rowsAligned X and ySeeded train rowsReserved test rows

Making a reproducible split

Protect evaluation rows before fitting.

12–18 min · 3 practices
F08Core · teaching
Predicting classes: Feature space, Class A or B, New row labelFeature spaceClass A or BNew row label

Predicting classes

Distinguish class labels from quantities.

12–18 min · 3 practices
F09Core · teaching
Preserving class representation: A and B classes, Training mix, Test mixA and B classesTraining mixTest mix

Preserving class representation

Apply and explain stratification.

12–18 min · 3 practices
F10Core · teaching
Could we know this at prediction time?: Known at decision, Prediction time, Later outcomeKnown at decisionPrediction timeLater outcomeNOW

Could we know this at prediction time?

Recognise leakage and context-dependent shortcuts.

12–18 min · 3 practices
F-R2Core · review
Honest evaluation retrieval: Training evidence, Final evidence, Keep them apartTraining evidenceFinal evidenceKeep them apart

Honest evaluation retrieval

Retrieve earlier concepts before combining them.

15–20 min · 3 practices
F-K1Core · checkpoint
ML Foundations checkpoint: Define X and y, Split, fit, predict, Read the errorDefine Xand ySplit, fit,predictRead theerror

ML Foundations checkpoint

Define X/y, split, fit, predict and evaluate a line.

25–35 min · 1 practices

04 / First complete workflows