GaussFit, A System for Least Squares and Robust Estimation
**Previous
abstract** **Next
abstract**

**Session 28 -- Software and Catalogs**
*Display presentation, Tuesday, 31, 1994, 9:20-6:30*

## [28.04] GaussFit, A System for Least Squares and Robust Estimation

*B. McArthur, W. Jefferys, J. McCartney (Department of Astronomy, University of Texas at Austin)*
GaussFit was developed as a platform to facilitate the solution of least squares
and robust estimation problems for astrometric data reduction with data from
the NASA Hubble Space Telescope. An environment where astrometric models
could easily and quickly be written, tested and modified was required.
GaussFit is capable of handling situations that arise often enough to be of
practical interest, but which have usually been ignored because they are not
well understood by many users. It provides an easy and natural way to
formulate general nonlinear problems; problems where the observation
equations (equations of condition) contain more than one observation (the
errors-in-variables case); problems with correlated observations; problems
where exact constraints among parameters must be enforced. Certain robust
estimation methods that generalize least squares to non-euclidian metrics and
provide greater immunity against RoutliersS than does the classical least
squares method are available. GaussFit runs both under UNIX and VMS and
there is a Macintosh version.

**Tuesday
program listing**