Regression vs Classification

These are the two big types of jobs supervised learning can do: guess a number, or guess a category.

Imagine it like this: "How many candies are in this jar?" is a number question (regression). "Is this candy chocolate or fruit?" is a category question (classification).

Regression examples:

• Predict tomorrow's temperature
• Predict how long a delivery will take

Classification examples:

• Decide whether a photo shows a cat or a dog
• Decide whether an email is spam

So? Before building a model, always ask yourself: "Am I guessing a number, or picking a category?" That tells you which tools to use.