Expand description
Penalized smoothing spline (ISL ch 7.5) — least squares with a roughness penalty that shrinks the wiggly (knot) part of the fit.
Minimise ‖y − Bβ‖² + λ·βᵀPβ, where B is the truncated-power spline basis
(shared with super::RegressionSpline) and P penalizes only the
truncated-power (knot) coefficients — the part that controls smoothness. λ = 0
reproduces the (interpolating) regression spline; large λ shrinks the knot
terms toward a global polynomial (a smooth fit). The penalized normal equations
(BᵀB + λP)β = Bᵀy are solved with linear_algebra::cholesky (no new solver).
Kernel-class DenseLinear.
Structs§
- Smoothing
Spline - A fitted smoothing spline.