🧊 k-NN Feature Space: 3D Decision Boundary (Blender)
3D k-nearest neighbours in a Blender-modelled feature cube: three features, five distance metrics with the true L1/L2/L-infinity distance balls, vote weights, a tappable decision slice, leave-one-out accuracy and four numerical checks.
⚙ Under the hood
3D k-nearest neighbours in a Blender-modelled feature cube: three features, five distance metrics with the true L1/L2/L-infinity distance balls, vote weights, a tappable decision slice, leave-one-out accuracy and four numerical checks.
KNNk-nearest neighborsmachine learningclassificationdecision boundaryleave-one-outThree.jsBlender
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install