Model Families and Honest Comparison
Learners can compare bounded candidate models against a dummy baseline and choose the simplest adequate model.
Executable courseRead the lesson pages, run the browser practice notebook, and keep the downloadable fallback pack if you want a local copy.
Module sequence
- What model comparison is allowed to claimUnit 1 · 20-35 minutes
- Dummy baselines and simple baselinesUnit 2 · 20-35 minutes
- Linear models and regularisationUnit 3 · 20-35 minutes
- Nearest neighbours and the role of distanceUnit 4 · 20-35 minutes
- Decision trees and readable rulesUnit 5 · 20-35 minutes
- Ensemble intuition: random forests and gradient boostingUnit 6 · 20-35 minutes
- Hyperparameters and bounded searchUnit 7 · 20-35 minutes
- Cross-validated model selectionUnit 8 · 20-35 minutes
- When to stop: choosing the simplest adequate modelUnit 9 · 20-35 minutes
- Bridge to Applied Machine Learning AlgorithmsUnit 10 · 20-35 minutes
- Project step: choose a final candidate honestlyUnit 11 · 20-35 minutes
- Activity: Compare Model Families HonestlyPractice activity · 45-75 minutes
- Module 08 Assessment: Model Comparison CheckModule check · 35-50 minutes