Download the notebook

This is the hands-on starter for Session 8. It is a fill-in notebook: the scaffolding is there, and you write the lines marked # TODO. Open it in Colab (button above) or download and run it locally; you only need numpy, matplotlib, and scikit-learn.

By the end you will be able to:

• Load a real dataset and inspect its class balance before fitting anything
• Make an honest, stratified train/test split and fit logistic regression
• Read the confusion matrix, precision, and recall instead of trusting accuracy
• Move the decision threshold and see precision trade against recall
• Extend the same idea to many classes with softmax, and read a per-class report

The lab starts with the Breast Cancer Wisconsin dataset (binary), then generalises to a ten-class problem with the handwritten digits dataset. A final section points you at other multi-class datasets to try (wine, Palmer penguins, iris).