Regression vs classification

• Regression: predict a number
• Classification: predict a label

Often classification models produce probabilities:

• P(class = 1 | X)

Then apply a threshold to decide the label.

Decision boundary intuition

A classifier learns a boundary that separates classes.

  flowchart LR
  X[Features] --> M[Classifier]
  M --> P[Probability scores]
  P --> T[Threshold]
  T --> Y[Predicted label]

Binary vs multiclass

• binary: 2 classes (0/1)
• multiclass: 3+ classes

Common pitfalls

• class imbalance
• wrong metric choice
• leakage via preprocessing

Mini-checkpoint

Choose a problem and write:

• classes (labels)
• cost of false negatives vs false positives
• best metric for the situation

🧪 Try It Yourself

Exercise 1 – Train-Test Split

Exercise 2 – Fit a Linear Model

Exercise 3 – Evaluate with MSE