Module 5 · Classical ML
Teach and adapt Module 5
Link the entry page directly. Its reading-site route is /lessons/05-regression/ relative to the served site base. For repository links, select a tested commit or release.
Support independent learners
Start with the forecast contract and hand arithmetic, then the stage scripts, commented notebook, and memo practice. Learners need functions, paths, and dataframe selection; the foundation bridge supplies the numerical concepts. Pilot pacing.
The script/text route can demonstrate training and interpretation without an inaccessible notebook UI or screenshot requirement. Notebook-interface operation is a separate skill if locally required. Discussion has a solo equivalent; GenAI is optional and local policies govern submissions.
Adapt the data and model together
Retain the UCI attribution/provenance. The synthetic fixture is an openly labeled offline alternative, with limited claims. For another dataset, specify unit, target, origin, horizon, feature availability, rights/privacy, and split before fitting. Repeated users/sites may need group boundaries; time series may need gaps or other availability modeling.
When changing features/models, update the contract, pipeline/exporter, inference checks, feedback, and tests. Numeric JSON supports only the narrow model kinds implemented here. Public historical test labels cannot become a blind benchmark through a script guard.
Alignment opportunities
| Outcome | Original register | Course alignment | Advisory ABET opportunity | Evidence |
|---|---|---|---|---|
| M05.1 | O05.1 | CLO1 support; CLO4 support | A1 | Equation/residual/loss |
| M05.2 | O05.2 | CLO2; CLO4 | A2 | Pipeline and baselines |
| M05.3 | O05.3 | CLO2; CLO4 | A1, A2 | Feature timing audit |
| M05.4 | O05.4; explicit evaluation added | CLO4; CLO8 | A1, A2 | Split, frozen choice, errors |
| M05.5 | O05.4 | CLO2; CLO8 | A3 | Reproducibility and memo |
These labels follow the separate curriculum planning register, not universal course identifiers. A1/A2/A3 are advisory computing-outcome opportunities for problem analysis, solution evaluation, and communication. Introductory artifacts do not prove master's-level attainment or accreditation compliance. Confirm adopted criteria before formal program use.
Verification and accessibility
CI exercises real-data download/workflow, synthetic behavior tests, notebook code, and site/report generation. Consult DEVELOPMENT.md and commit-specific results. Notebook code checks are not interface usability tests.
Reports supply named values, scoped/captioned tables, CSV/text evidence, and descriptions alongside optional plots. Consult the accessibility evidence log for human interface-review coverage. Provide a private help route and test actual learner tasks before requiring an interface.
Teaching prototype: explanation, practice and feedback
This module now has an intermediate concepts page and three independently completed functions. Use the pilot protocol to observe understanding and support needs. Map learner artifacts with the practice evidence table. Visible code checks and solution execution are technical evidence; review the learner's reasoning and transfer separately. Preserve the existing outcomes and official assessment rules.