Abstract
The smoothing spline estimator minimizes a penalized least squares criterion over functions with square integrable
mth-order partial derivatives. We prove it converges almost surely and uniformly over the entire domain as the sample size increases.