3D Random Forest Visualization (engine)

⚙ Forest Controls

Number of trees 5

Max depth 3

Feature 1 0.5

Feature 2 0.5

📋 Live Readout

Trees5

Votes for Class 0-

Votes for Class 1-

Confidence-

Feature vector0.5, 0.5

NO PREDICTION YET

Drag to orbit · scroll to zoom · each tree votes with the same threshold-split logic as the 2D version

3D Random Forest Visualization

An orbitable Three.js forest of decision-stump trees. Each tree is built with the identical predict(features) logic as the 2D Random Forest Visualization: a random threshold split on one of two features, then a second threshold split into leaf class 0 or 1. Generate a new forest, tune the two feature sliders, and press Predict to watch each tree cast its vote — highlighted leaves and animated particles fly to the class-0/class-1 vote pillars, whose heights scale with the live vote tally, exactly mirroring the majority-vote rule of the 2D bar chart.