Week 6: Classification basics (kNN)

Learning Objectives

• 06A I can train and interpret k-NN classification models and explain how the choice of k affects overfitting and underfitting.
• 06B I can compute and interpret classification metrics — accuracy, precision, recall, and the confusion matrix.
• 06C I can split data into training and test sets to evaluate how well a model generalizes to unseen data.

Perspectival Reading

Reading: TBD

Reflection Questions

1. kNN defines similarity by distance in feature space — whose definition of similarity does that embed?
2. Classification assigns people or things to categories. What is the cost of a wrong assignment, and who bears it?
3. “Accuracy” treats all errors as equal. When is that assumption unjustified?

Materials

• Monday class
• Wednesday class