Home▸Algorithms & AI▸Bias-Variance Tradeoff: Polynomial Fits (2D)

Bias-Variance Tradeoff: Polynomial Fits (2D)

A 2D XY-plot companion to the 3D lab: least-squares polynomial curves are refit to several independently noisy resamples of the same hidden function, and you can watch bias² and variance trade off directly as model complexity, noise and sample size change.

Algorithms & AI2DModerate60 FPS📱 Mobile-adapted⇄ 3D version

This 2D companion draws the same least-squares polynomial fits as the 3D lab on a plain XY plot instead of a stacked-depth scene: several noisy resamples of a hidden sine-based function are each refit independently, the orange curves show how tightly (low degree) or loosely (high degree) they agree with one another, and the live bias²/variance/total-error readout updates instantly as you drag the complexity, noise, sample-size or resample-count sliders.

⚙ Under the hood

2D polynomial-regression bias-variance lab: refit resampled noisy training sets on a plain XY plot and watch bias squared and variance trade off as model complexity changes.

machine-learningregressionpolynomialresamplingmodel fitting

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install