Definition
A random forest builds many decision trees on bootstrapped samples, each considering a random subset of features at splits. Predictions are averaged (regression) or voted (classification).
It resists overfitting better than a single deep tree and needs relatively little preprocessing.
In simple terms
Instead of one opinionated flowchart, you ask hundreds of slightly different flowcharts and take the majority vote.
Where you see it
- Feature importance for tabular business problems.
- Baseline models before trying neural nets.
How it works
1.Bootstrap samples
Each tree sees a random draw of rows.
2.Random features
Splits consider a subset of columns.
3.Aggregate trees
Combine predictions for the final answer.
Why it matters
- Random forests are a reliable default for many structured-data ML tasks.
Often confused
Random forests work great on raw text and images.
They need tabular features; for text/images, neural nets or embeddings usually win.