AlgoPlus//machine learning / logistic-regression
Read the theory

Logistic Regression

A linear score squashed by a sigmoid into a class probability.

Loss
Iter
Legend
Class 1
Class 0
σ = 0.5 boundary
AI Tutor Workspace
In a nutshell
Logistic regression sorts things into two classes and reports how sure it is, not just yes or no. It scores each point by which side of a line it's on, then runs that score through the S-shaped sigmoid to get a probability between 0 and 1; training slides the line to make those probabilities match the real labels.
Ready
Press play to begin the cinematic walkthrough.
Draw a line, but instead of a hard yes/no, report how confident you are: the sigmoid turns 'distance from the line' into a probability that saturates at 0 and 1.
Key terms
Go deeper in the lesson
Read the full theory, intuition & complexity for Logistic Regression.