Learn AI · Beginner · 3 min

How decision trees help AI choose

Can a computer learn by asking yes-or-no questions?

Decision tree structure and flow

Data (examples)Root: first questionBranch: yes or noLeaf: final answerTree grows too deep

Start at the top. Answer each question. Follow the branch. Reach a leaf with the answer.

goes inthe AI workscomes outwatch out

In simple words

A decision tree is like a flowchart that asks questions to sort things. Each answer leads to a new question until you reach an answer.

The real definition

A decision tree is a supervised learning model that makes decisions by splitting data into groups using yes-or-no questions. It mimics how humans think through choices.

Like… Playing 20 questions

You guess what someone is thinking by asking yes-or-no questions. Each answer narrows down the options until you find it. The decision tree does exactly this with data.

But: Real 20 questions stops when you guess. Trees may ask too many questions and memorize instead of learning.

You see it every day

on your phone

Email spam filter

A tree asks: "Is the sender unknown?" Then "Does it ask for money?" The AI sorts email as spam or safe.

out in the world

Animal classifier

A tree asks: "Does it have feathers?" Then "Can it fly?" The AI tells you if it is a bird.

at work

Loan approval

A tree asks: "Is income above the limit?" Then "Is credit score good?" The AI decides yes or no.

Step by step

1
Gather training data
Collect examples with labels. Labels are the right answers the AI should learn.
2
Pick the best question
Find a yes-or-no question that splits the data into two clean groups.
3
Repeat for each branch
Ask more questions on each group. Stop when each leaf has only one type.
4
Stop at the right depth
Stop growing when leaves are pure or the tree gets too big.

Remember

Trees ask yes-or-no questions to sort dataEach question splits data into two groupsLeaves (endpoints) hold the final answerToo-deep trees memorize instead of learningFast and easy for humans to understand

Memory trick

Think of a real tree: root at top, branches split left and right, leaves at the bottom are where answers hide.

Words

Decision tree de-SIZ-un TREE: A model that sorts data by asking yes-or-no questions in order.
Root ROOT: The first question at the top of the tree.
Branch BRANCH: A path that follows a yes or no answer from a question.
Leaf LEAF: The endpoint where the tree gives the final answer.
Supervised learning SOO-per-VYZD LER-ning: Teaching AI using examples that already have correct answers labeled.
Split SPLIT: Dividing data into two groups using a yes-or-no question.
Depth DEPTH: How many questions the tree asks from root to leaf.
Overfitting OH-ver-FIT-ing: When a model memorizes examples instead of learning real patterns.

Check yourself

quick check

What does a decision tree ask at each step?

Hint

Think of how you play 20 questions.

quick check

A bank trains a tree to approve loans. The tree asks ten questions before deciding. What is the risk?

Hint

Deeper trees learn less. They memorize more.

think

Why is a decision tree easier to understand than other AI models?

Show a good answer

Because you can follow the path from root to leaf and see why it decided something. Each question and answer makes sense to humans.

Tell a friend

“A decision tree is an AI that learns by asking yes-or-no questions to sort things into groups.”