Activity: Train First Supervised Baselines

Activity ID: ML-M03-A01 Estimated active time: 55-75 minutes

Goal

Create a short baseline comparison note for one regression task and one classification task.

Inputs

• data/module_03_regression_baseline_v1.csv
• data/module_03_classification_baseline_v1.csv
• data/data-dictionary.md
• notebooks/module-03-worked.ipynb
• notebooks/module-03-practice.ipynb

Instructions

1. Open the data dictionary and confirm the two targets.
2. List the numeric feature columns recommended for Module 3.
3. Run the regression section of the worked notebook.
4. Record the median baseline MAE and linear regression MAE.
5. Inspect at least five actual vs predicted quiz scores.
6. Run the classification section of the worked notebook.
7. Record the most-frequent baseline accuracy and logistic regression accuracy.
8. Inspect at least five predicted probabilities.
9. Complete the practice notebook.
10. Write a short note explaining what the two first models suggest and what they do not prove.

Submission table

Quality checklist

• [ ] The target is not included in the feature table.
• [ ] Train/test split happens before model fitting.
• [ ] A dummy baseline is reported for both tasks.
• [ ] The candidate result is compared with the baseline.
• [ ] MAE is explained in quiz-score points.
• [ ] Accuracy is explained as a first metric, not the only metric.
• [ ] The explanation avoids saying the model is ready for real use.
• [ ] The note says the data is synthetic.

Submission format

Submit a completed Markdown table and a 4-6 sentence explanation. The notebook outputs can be used as evidence.

Safety note

Use only the supplied synthetic datasets. Do not add, upload, or paste real learner data.