Penalized hyperbolic-polynomial splines (Q2233247)

Splines and spline approximations using penality functions are especially useful tools in univariate approximation theory when univariate data are to be approximated, and when no interpolation property is needed. The methods using piecewise polynomial splines and penalty methods can be formulated via square (or otherwise, other \(L^p\)-norms are possible, \(p=1\) for instance, albeit rarely used) integrals over (second, for example) derivatives. Alternatively, derivatives can be replaced by discrete expressions, divided differences. In this article, this approach is persued but with piecewise exponential (times monomials) replacing piecewise polynomials. The polynomial case is of course a special case when the exponents have zero-parameters.