5.0 Study Guide and Task Sheet: Machine Learning
The main study route is now in Chapter 5 entry. Use this page only as a quick checklist while you practice.
One-Line Mental Model
define task -> split data -> train baseline -> evaluate -> inspect errors -> improve
If you do not know which model to use, start with a baseline.
Practice Checklist
| Check | Evidence |
|---|---|
| I can define the task type | problem note |
| I can split data without leakage | train/test split note |
| I can train a dummy baseline and one real model | baseline comparison |
| I can choose a metric for the task | metric note |
| I can inspect errors | error samples |
| I can finish the evidence-pack workshop | ml_workshop_run/ |
Evidence Rubric
| Artifact | It should answer |
|---|---|
| Problem note | What is the task type, and what counts as success? |
| Split note | How did you keep test data away from training? |
| Baseline comparison | What is the minimum score to beat? |
| Metric note | Why does this metric match the goal better than plain accuracy? |
| Error note | Which mistakes matter most, and what feature or label issue might explain them? |
Ready To Continue
Continue to Chapter 6 when one tabular project includes a baseline, a real model, metrics, error analysis, and a README that another person can rerun.