Random Forest Voting
A random forest averages many decision trees, each trained on a bootstrap resample with randomised split thresholds, so the trees’ individual errors vote each other away. Set the number of trees, their maximum depth and the bootstrap fraction, press Play, and trees are added one at a time: the heatmap softens as votes accumulate, showing the fraction of trees choosing each class at every location, while the curve below tracks test accuracy against ensemble size. Switch to the single-tree view or the side-by-side comparison to see one jagged tree against the smooth ensemble, or overlay a few individual tree boundaries to see the diversity being averaged. Click to add points; a fixed 70/30 split scores accuracy.
Runs 100% in your browser — models are trained and computed locally on your device.
Read the full guide to this tool
Notes
- Bagging: every tree sees a bootstrap sample drawn with replacement, which leaves out roughly 37% of the training points per tree.
- Random split thresholds decorrelate the trees; averaging n weakly correlated trees cuts the variance of the vote roughly like 1/n.
- Adding more trees never overfits the way deepening one tree does — the accuracy curve climbs and then flattens to a plateau.
- Runs 100% in your browser — models are trained and computed locally on your device.
Related tools
Decision Tree Builder
Grow a CART decision tree split by split and watch the partition map and tree diagram build together.
Function Sketch Finder
Sketch a curve on a zoomable coordinate grid and search dozens of function families for the equations that fit it best.
Linear Regression Playground
Place points, watch the least-squares line follow, and see gradient descent crawl across the loss surface.
Overfitting Sandbox
Crank a polynomial’s degree on noisy data and watch the train/test error curves split apart.
Logistic Regression Playground
Paint two classes of points and watch logistic regression carve out its probability boundary by gradient descent.
Naive Bayes Classifier Demo
Score your own message with a word-count spam filter, or watch Gaussian naive Bayes classify 2D points.