Constrained smoothing splines revisited

Research Report SRR97-008

Constrained smoothing splines revisited

Berwin A. Turlach

Abstract: In some regression settings one would like to combine the flexibility of nonparametric smoothing with some prior knowledge about the regression curve. Such prior knowledge may come from a physical or economic theory, leading to shape constraints such as the underlying regression curve being positive, monotone or convex-concave. We propose a new method for calculating smoothing splines that fulfill these kind of constraints. Our approach leads to a quadratic programming problem and the infinite number of constraints are replaced by a finite number of constraints that are chosen adaptively. We show that the resulting problem can be solved using the algorithm of Goldfarb and Idnani (1982, 1983) and illustrate our method on several real data sets.

This service is maintained by the Mathematical Sciences Institute (MSI)
Comments to webmaster@maths.anu.edu.au URL: http://wwwmaths.anu.edu.au/