AI Glossary Term
What does “Decision Tree” mean?
A model that makes predictions through a sequence of if-then feature splits.
How Decision Tree works
Example
A bank might use a tree that first asks about income, then about existing debt, and ends in approve, review, or decline. Each path is a readable list of conditions, so staff can see why a case ended where it did.
Why it matters
Trees are among the easiest models to read, which matters when decisions must be explained. Groups of trees, such as random forests, are also strong on table data.
Common questions
How does a tree choose its questions?
It tries splits on the data and picks those that best separate the outcomes, using measures such as Gini impurity or information gain.
Why do trees overfit?
A deep tree can create a branch for nearly every training example, memorizing noise. Limiting depth or pruning helps.
What is a random forest?
An ensemble of many trees trained on different samples of the data whose votes are combined for a more stable prediction.