LSQFT: a nonlinear least squares data fitting subroutine suitable for minicomputers. [In FORTRAN for Hewlett Packard 1000]
The FORTRAN subroutine LSQFT is described which uses the Gauss method to fit nonlinear functions to data. This particular program is suitable for use on a minicomputer that is utilized for real-time data acquisition and analysis. A sample problem is described that requires a nine-parameter nonlinear fit to 200 data points. This problem serves both to illustrate the use of LSQFT and to compare its results with TJMAR1, a nonlinear parameter estimation program available in the Sandia MATHLIB, for the CDC 6600/7600 computers. The results of this comparison fit are identical to the significance of the input convergence criteria. The output of LSQFT includes the standard deviation of each fit parameter, which is useful for analysis of the significance of the parameters based on the input data or for evaluation of functional models used to fit to the data. Examples are given of the types of graphical output that can be generated from the results of LSQFT. The program is documented and listed in its entirety, and a user's guide is provided. 10 figures.
- Research Organization:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- DOE Contract Number:
- AC04-76DP00789
- OSTI ID:
- 5254362
- Report Number(s):
- SAND-80-0496
- Country of Publication:
- United States
- Language:
- English
Similar Records
EXPALS: FORTRAN code for exponential approximation by least squares. [For CDC 6600 or 7600]
A data analysis environment for HP (Hewlett-Packard) 1000 computers
Related Subjects
COMPUTER CODES
L CODES
LEAST SQUARE FIT
COMPARATIVE EVALUATIONS
DATA ACQUISITION
DATA ANALYSIS
DATA COVARIANCES
FORTRAN
HP COMPUTERS
NONLINEAR PROBLEMS
PERFORMANCE
REAL TIME SYSTEMS
COMPUTERS
MAXIMUM-LIKELIHOOD FIT
NUMERICAL SOLUTION
PROGRAMMING LANGUAGES
990200* - Mathematics & Computers