# Use a pipeline and a genuinely held-out set Open Practice / Measured machine learning / curriculum 1.0 Estimated guided practice: 4h estimated. This estimate has not been learner-tested. Original practice material · CC BY 4.0 · Non-degree and non-accredited. ## What you will build Ship a versioned experiment bundle with inputs, tests and a reproducible handoff. ## Before you start - Python + statistics; inspect the worked fixture before beginning. - A text editor/spreadsheet for the manual route. Hosted Python/Jupyter for code exercises; applicable dependencies are in requirements.txt. - For actual AI calls, a legitimate available cloud service and authorised learner credentials; no particular provider or country is mandatory. ## Learning outcomes - Write the split and allowed tuning procedure before fitting. - Compare against majority class and transparent rules. - Use a pipeline so training parameters stay within the correct partition. - A four-point synthetic success checks wiring, not field reliability. - Keep a reproducible result and disclose safety, rights and regional access limits. ## Activity 1: Reserve a final holdout Estimated time: 35 minutes. ### Outcomes - Write the split and allowed tuning procedure before fitting. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. Write the split and allowed tuning procedure before fitting. Inspect the supplied input before using any tool. Write the decision rules in your own words and calculate one small result independently. The reference output is an answer key for this deliberately bounded fixture, not proof of general capability. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. For learners across regions underserved or excluded by mainstream providers: verify country eligibility, language, account/identity requirements, data terms, cost, quota and connectivity before choosing a cloud service. Use only legitimate available alternatives; never bypass restrictions or invent a country/account. No service access? Complete the labelled manual/fixture route without claiming a live AI run. ### Worked example ```text SOURCE (authored fixture): { "train_x": [ [ 0 ], [ 1 ], [ 2 ], [ 8 ], [ 9 ], [ 10 ] ], "train_y": [ 0, 0, 0, 1, 1, 1 ], "test_x": [ [ 0.5 ], [ 1.5 ], [ 8.5 ], [ 9.5 ] ], "test_y": [ 0, 0, 1, 1 ] } INDEPENDENT FIXTURE TARGET: { "predictions": [ 0, 0, 1, 1 ], "correct": 4, "test_count": 4 } ``` ### Guided activity 1. Write the split and allowed tuning procedure before fitting. 2. Open input.json and expected.json from the practice pack; keep an unmodified original. 3. Record a manual prediction and the relevant rule before consulting the answer key. 4. Perform one small change and predict its effect before rerunning. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Reserve a final holdout — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Move a test point near the class boundary: inspect uncertainty instead of demanding a perfect score. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## Activity 2: Fit the simplest baseline Estimated time: 35 minutes. ### Outcomes - Compare against majority class and transparent rules. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. Compare against majority class and transparent rules. Establish a baseline before changing prompts or code. Keep the original input immutable. Record expected values, observed values and whether each comparison passed. If the expected file conflicts with a justified policy, investigate rather than forcing a convenient match. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. ### Worked example ```text MANUAL BASELINE / compare every named field { "predictions": [ 0, 0, 1, 1 ], "correct": 4, "test_count": 4 } Keep a separate observed-result column; the answer key is not a model response. ``` ### Guided activity 1. Compare against majority class and transparent rules. 2. Open input.json and expected.json from the practice pack; keep an unmodified original. 3. Record expected and observed values separately; explain every mismatch. 4. Perform one small change and predict its effect before rerunning. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Fit the simplest baseline — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Scale on all rows before splitting: identify leakage. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## Activity 3: Compose preprocessing and model Estimated time: 35 minutes. ### Outcomes - Use a pipeline so training parameters stay within the correct partition. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. Use a pipeline so training parameters stay within the correct partition. Build one inspectable transformation or bounded AI prompt. Keep calculation and validation deterministic where possible. If you make a permitted hosted call, save the actual response separately from the authored examples and verify its material claims; the supplied reference never simulates a live model. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. ### Worked example ```text PROJECT CONTRACT: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. HOSTED REFERENCE: open hosted-lab.ipynb or run reference.py with all fixture files together. EXPECTED CONSTRUCTED RESULT: { "predictions": [ 0, 0, 1, 1 ], "correct": 4, "test_count": 4 } ``` ### Guided activity 1. Use a pipeline so training parameters stay within the correct partition. 2. Open hosted-lab.ipynb and the reference code from the practice pack; keep an unmodified original. 3. Record expected and observed values separately; explain every mismatch. 4. Perform Try another seed or a harder dataset: keep the new result even if performance drops. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Compose preprocessing and model — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Try another seed or a harder dataset: keep the new result even if performance drops. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## Activity 4: Inspect predictions and errors Estimated time: 35 minutes. ### Outcomes - A four-point synthetic success checks wiring, not field reliability. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. A four-point synthetic success checks wiring, not field reliability. Change one assumption at a time and predict the result first. Retain the failure, the revision and the rerun evidence. A failure is useful information: hiding a bad output or editing a test target after seeing the answer destroys the value of the experiment. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. ### Worked example ```text CHANGE TESTS — predict before running: 1. Move a test point near the class boundary: inspect uncertainty instead of demanding a perfect score. 2. Scale on all rows before splitting: identify leakage. 3. Try another seed or a harder dataset: keep the new result even if performance drops. Keep the original expected.json unchanged. ``` ### Guided activity 1. A four-point synthetic success checks wiring, not field reliability. 2. Open failure-tests.md from the practice pack; keep an unmodified original. 3. Record expected and observed values separately; explain every mismatch. 4. Perform Move a test point near the class boundary: inspect uncertainty instead of demanding a perfect score. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Inspect predictions and errors — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Move a test point near the class boundary: inspect uncertainty instead of demanding a perfect score. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## Activity 5: Compare validation schemes Estimated time: 35 minutes. ### Outcomes - Explain when random versus time-based splits are appropriate. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. Explain when random versus time-based splits are appropriate. Review what the fixture cannot establish. Explain who owns the decision, where human approval is needed and which access or professional constraints remain. Do not describe a small educational success as production readiness, certification, legal compliance or a guarantee. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. For learners across regions underserved or excluded by mainstream providers: verify country eligibility, language, account/identity requirements, data terms, cost, quota and connectivity before choosing a cloud service. Use only legitimate available alternatives; never bypass restrictions or invent a country/account. No service access? Complete the labelled manual/fixture route without claiming a live AI run. ### Worked example ```text REVIEW BOUNDARY: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. ACCESS / NOT TESTED: unavailable provider, unsupported language and absent external feedback must remain explicit. ``` ### Guided activity 1. Explain when random versus time-based splits are appropriate. 2. Open README.md and your evidence log from the practice pack; keep an unmodified original. 3. Record the human decision owner, limitations and any untested service access. 4. Perform one small change and predict its effect before rerunning. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Compare validation schemes — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Scale on all rows before splitting: identify leakage. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## Activity 6: Ship a versioned experiment bundle Estimated time: 65 minutes. ### Outcomes - Include parameters, held-out predictions, software versions and a model card. - Explain and retain a testable result and its limits. A library pipeline fits transformations and a predictor together, reducing accidental leakage. Cross-validation belongs inside the training partition. A final holdout is a one-time audit, not a tuning dashboard. Include parameters, held-out predictions, software versions and a model card. Deliver a reproducible bundle rather than a screenshot alone. Include inputs, rules, output, failure tests, a short README and residual limitations. Ask another person to follow the instructions if possible; self-review and actual external review must be labelled separately. Policy for this project: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. ### Worked example ```text HANDOFF MANIFEST: input.json — immutable source expected.json — fixture answer key actual.json — result actually observed evidence-log.csv — baseline plus three changes README.md — reproduction and access route limitations.md — unresolved failures, rights, approvals and scope Never bundle a credential. ``` ### Guided activity 1. Include parameters, held-out predictions, software versions and a model card. 2. Open README.md and your evidence log from the practice pack; keep an unmodified original. 3. Record the human decision owner, limitations and any untested service access. 4. Perform one small change and predict its effect before rerunning. 5. Keep the result and your reasoning in the lesson notes; do not mark a simulated run as live. ### Expected evidence - Ship a versioned experiment bundle — a saved, inspectable artefact tied to the supplied rule. - A baseline or change-test record with expected/observed results and a limitation. ### Common mistakes - Try another seed or a harder dataset: keep the new result even if performance drops. - Treating a supplied answer key, an authored example or self-review as independent field validation. ### Check your understanding Which exact input, rule and expected value would let another learner falsify your conclusion in this activity, and what remains untested? ### Offline route Use the supplied fictional JSON, expected results and paper/text evidence log without a provider account. Manual or constructed completion is labelled; it is not a live cloud integration or trained-model run. ## End-of-path review These are self-check questions, not verified examinations. ### Question 1 For the supplied fixture, what is the expected predictions? 1. [0, 0, 1, 1] 2. It cannot be checked against the supplied input. 3. Whatever value the model reports most confidently. Review answer: 1. The independent answer key only validates this constructed fixture. Preserve the source, rerun counterexamples, and disclose what was not tested. ### Question 2 How should you handle the project rule: This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. 1. Silently change the rule to fit an attractive output. 2. State the rule, test it, and record exceptions or unresolved failures. 3. Treat model confidence as proof. Review answer: 2. Rules, source provenance and counterexamples make an output inspectable; an attractive result does not override policy. ### Question 3 If mainstream providers do not serve your region, what is the honest route? 1. Use a false country or borrowed account. 2. Claim the supplied fixture was a real API run. 3. Check legitimate available alternatives, or label the manual/fixture route and its limits. Review answer: 3. Access includes country, identity, language, cost and data terms. Do not bypass restrictions or invent a live result. ## Capstone review rubric Score each criterion 0 (missing), 1 (partial), or 2 (demonstrated). A self-review pass needs at least 8/10 and 2/2 on the critical safety/limits criterion. This is learner practice feedback, not verified assessment. ### Defined task and preserved input Original input and the project policy are retained; the actual question is clear. ### Checked baseline The supplied baseline reconciles field by field; unsupported live-run claims are absent. ### Three failure/change tests All three specified changes have predictions, observed results and explanations. ### Repeatable handoff Another person can follow the README; supporting practice evidence and unresolved limits are included. ### Safety, rights and access honesty [critical] This tiny separable synthetic fixture is a wiring test only. Fit StandardScaler and LogisticRegression on six rows; predict four held-out rows. Record versions and do not generalise the resulting score. Regional access route is legitimate; credentials and private data are absent; manual/constructed work is clearly labelled. ## References - [Scikit-learn Pipeline reference](https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.Pipeline.html) ## Editorial boundary Original authored practice material; deterministic fixtures build-tested. Not independently expert-reviewed or learner-validated. Time estimates unvalidated. Use fictional/redacted data and your own authorised provider credentials. Never include credentials in this pack or public work.