Regression: Predicting a Number
A free lesson from Swipe AI · Machine Learning Fundamentals, level 2 · updated October 2026
By the Swipe team, founded by Colin · how we make lessons
Regression shares supervised learning's core setup — labeled training examples — but the output is a number along a continuous range rather than a category. That changes what 'correct' even means: a classifier is either right or wrong about the category, but a regression prediction can be closer or further from the true number.
Take a food delivery app predicting delivery time: the inputs might be distance, current traffic, and time of day, and the output is a number of minutes. If the true delivery took 34 minutes and the model predicted 31, that's a small, forgivable miss — nowhere near as bad as predicting 5 minutes or 90.
The same input/output/training-examples shape from Level 1 still applies: for house price prediction, the inputs are features like square footage, location, and number of bedrooms; the output is a predicted price; and the training examples are real past sales where the true final price is already known.
Some things can be framed either way depending on what you actually need: predicting a movie rating as 'exactly 7.3 out of 10' is regression, but predicting it as 'good, okay, or bad' is classification. The right framing depends on whether you need a precise number or just a useful bucket.
Why can a regression prediction be "closer" or "further" from the truth in a way a classification prediction cannot?
Predicting the exact number of minutes a delivery will take, based on distance, traffic, and time of day, is an example of ____.
The same real-world question can sometimes be framed as either a classification or a regression problem, depending on whether you need a precise number or a useful category.
Classification vs. Regression
10 questions · pass with 60% — just like in the app.
Which best describes a classification task?
Which of these is a regression problem?
A classification model can output a category that never appeared in its training examples.
A supervised learning task where the output is a number along a continuous range is called ____.
What is multi-class classification?
A regression prediction can be closer or further from the true value, unlike a classification prediction, which is simply right or wrong about the category.
Most classifiers estimate how likely an input belongs to each category, then return the category with the highest ____.
Predicting a movie rating as "good, okay, or bad" instead of "7.3 out of 10" reframes the problem from regression into what?
The same real-world question can sometimes be framed as either classification or regression, depending on whether a precise number or a useful category is needed.
In a house-price prediction system, the price being predicted is the ____, and since it is a continuous number, this is a regression problem.
Free ai tools
Keep your streak going in Swipe AI
Swipe through every level of Machine Learning Fundamentals on your iPhone — with XP, stars, streaks, a personal advisor and the full track beyond these free levels.
Get the Swipe apps →