Bias-Variance Tradeoff Interactive Visualization

Polynomial Degree: 3

Noise Level: 0.5 Training Samples: 20

0.00

Systematic error
(underfitting)

0.00

Sensitivity to data
(overfitting)

0.00

Bias² + Variance
+ Noise

Bias: Error from wrong assumptions. High bias → underfitting (model too simple).
Variance: Error from sensitivity to training data. High variance → overfitting (model too complex).
Optimal Model: Balances bias and variance, minimizing total error.
Sweet Spot: Usually polynomial degree 2-4 for this data.
Experiment: Adjust complexity slider to see how bias and variance change.

⚖️ Bias-Variance Tradeoff