Logistic Regression

Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary

A classification method, despite its name: it fits a sigmoid curve to data and outputs the probability that an input belongs to one of two classes.

What is Logistic Regression?

Logistic regression is a classification method that predicts the probability an input belongs to one of two classes – despite a name that suggests it predicts continuous values. The name survives from statistics, where the technique grew out of regression analysis; the output, though, is a probability between 0 and 1 that gets converted into a class label.

If you need to predict a number, use regression. If you need to predict a category, logistic regression is one of the first tools to reach for.

How Logistic Regression Works

The model computes a weighted sum of the input features, then passes it through the sigmoid function, which squashes any number into the range 0 to 1:

\[ p = \sigma(w \cdot x + b), \qquad \sigma(z) = \frac{1}{1 + e^{-z}} \]

Training adjusts the weights until the predicted probabilities match the labels in the training data. A threshold, usually 0.5, then turns each probability into a decision: above it, one class; below it, the other.

Each weight has a readable meaning: it states how strongly its feature pushes the odds up or down. That is why logistic regression persists in regulated fields, where a lender or hospital must explain every automated decision.

Example of Logistic Regression

A bank builds a default-risk model from past loans. Features include income, existing debt, and payment history; the label records whether each borrower defaulted.

Training finds that a high debt-to-income ratio raises the weighted sum and past late payments raise it further, while a long clean record lowers it. A new applicant’s numbers produce a sum of 1.2, and the sigmoid turns that into a default probability of 0.77.

The bank auto-approves only below a 0.2 default probability, so this application routes to a human reviewer. Because every weight is visible, the bank states exactly which factors drove the score when fair-lending rules require an explanation.

Related AI terms: Classification · Sigmoid Function · Regression Analysis · Naive Bayes

Did you like the Logistic Regression gist?

Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.

Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About

Read the Governor's Letter

Stay ahead with Governor's Letter, the newsletter delivering expert insights, AI updates, and curated knowledge directly to your inbox.

By subscribing to the Governor's Letter, you consent to receive emails from AI Guv.
We respect your privacy - read our Privacy Policy to learn how we protect your information.

Browse All AI Terms A–Z

Every term in the dictionary, in alphabetical order. Jump to a letter or scroll the full list.

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z