Random Forest
Combine many decision trees to build a stronger random forest model.
A random forest combines many decision trees to make more stable predictions than a single tree alone.
Goal
Train a random forest and compare it to a single decision tree.
How Random Forests Work
Each tree in the forest is trained on a random subset of the data and features. The final prediction comes from voting across all trees, which reduces overfitting and improves generalization.
When to Use Random Forests
Random forests work well on structured tabular data and are a strong baseline before trying more complex models.
Interactive
Random Forest Voting Demo
Toggle each tree's prediction and see how majority voting determines the forest's final class.
Tree 1
Class 0
Tree 2
Class 1
Tree 3
Class 0
Tree 4
Class 1
Tree 5
Class 0
Tree 6
Class 1
Votes for Class 0
3
Votes for Class 1
3
Forest prediction
Tie
Majority vote
Each tree votes for Class 0 or Class 1. The random forest chooses the class with the most votes. If votes are equal, the result is a tie.
Tree votes: [0, 1, 0, 1, 0, 1] → Tie
Practice
Try It Yourself
Open the practice lab to complete the starter code in the notebook.
Knowledge Check
Quick Quiz
How does a random forest make a final prediction?
Summary
Key Takeaways
- A random forest combines many decision trees into one ensemble model.
- Voting across trees often produces more stable predictions.
- Random forests are a strong baseline for tabular classification problems.