Name: Regularization Methods Lab
Author: MySimulator

Model

Polynomial degree: 9 L2 regularization strength (λ): 0.00

A high-degree polynomial has enough flexibility to snake through every noisy training point exactly, which usually means it captures noise, not signal. L2 regularization penalizes large coefficients, forcing the fit to stay smoother even when the model has plenty of raw flexibility.

Error

Training error (RMSE)—

Test error (RMSE)—

Fit versus true function (dots = training data)