Logistic Regression

← Back to the article

Logistic regression turns clues into a chance, then chooses yes or no.

What it does

Logistic regression turns clues into a chance, then chooses yes or no.

Sigmoid: The S-shaped line • Binary: A choice between two things

How it works

  • Add up the clues.
  • Turn the score into a percent chance.
  • If the chance is high enough, answer yes. Otherwise, answer no.

Uses and applications

  • An email app can ask, “Is this spam or not spam?”
  • A school app can predict pass or fail from practice scores.
  • A shop can guess whether a visitor may buy something today.

In short

It is used for two-choice decisions, like spam or not spam.

Related methods

  • Linear Regression: Shares the weighted-sum idea, but predicts a number instead of a chance.
  • Naive Bayes: Another way to turn evidence into a probability for a class.
  • Neural Network: A single neuron with a sigmoid output computes the same kind of chance.

Back to Prediction and classification · All topics

The interactive demonstration and the Python walkthrough for this topic need JavaScript.