“…Therefore, the derivative of the smoothing spline at the knots satisfies the matrix equation where m = ( m 0 , m 1 ,…, m N ) T is the vector of slopes, T is the tridiagonal matrix and Note that the diagonal entries of T are all positive and T is strictly diagonally dominant, meaning that in each and every row the sum of the absolute values of the off‐diagonal elements is less than the absolute value of the diagonal element. Consequently, the symmetric tridiagonal matrix T is positive definite [see, e.g., Demmel , , Theorem 2.9; Andelić and da Fonseca , , Theorem 1.2]. The most efficient numerical method of solving the tridiagonal system T m = u , where u = D f , is based on the Cholesky decomposition T = R T R , where, because T is a symmetric positive definite tridiagonal matrix, R is an upper triangular matrix with all zeros above the first superdiagonal [ Schwarz , ; Golub and Van Loan , ].…”