Constrained least squares curve fitting to discrete data using B-splines: a user's guide. [FC for CDC 6600]
A subprogram, FC(), was developed that allows a user to fit discrete data in a weighted least-squares sense by using piece-wise polynomial functions. The piece-wise polynomial functions are represented by B-splines on a given set of knots. An existing package of codes, developed by C.W. de Boor, is used to evaluate the B-splines. The code FC() contains some important new features. In addition to the least-squares fitting of the data, a variety of constraints can be imposed on the fitted curve or its derivatives. These are equality, inequality, and perodic constraints at a discrete, user-specified set of points. This feature allows one to achieve desired qualitative properties in the resulting curve. Another feature of this code is the capability of dealing with large data sets. This is accomplished by the user's first accumulating the Q-R decomposition of the least-squares equations, and then passing the upper triangular factor to FC(). The computation of the residual variance function can be done by use of an additional function subprogram, CV(), which is provided. 3 figures, 2 tables.
- Research Organization:
- Sandia Labs., Albuquerque, NM (USA)
- DOE Contract Number:
- EY-76-C-04-0789
- OSTI ID:
- 6189927
- Report Number(s):
- SAND-78-1291
- Country of Publication:
- United States
- Language:
- English
Similar Records
FC,LSEI,WNNLS. Constrained Least Squares Fit
Quadrature subroutines for splines and B-splines. [BFQUAD, BSQUAD, PFQUAD, and PPQUAD]