GFIT - generalized quadratic approximation of functions under constraints

✍ Scribed by V.B. Zlokazov

• Publisher: Elsevier Science
• Year: 1989
• Tongue: English
• Weight: 635 KB
• Volume: 54
• Category: Article
• ISSN: 0010-4655
• DOI: 10.1016/0010-4655(89)90097-0

✦ Synopsis

A method and a program are described for the minimization of a function of the type E~w(x){y(x)-f(x, p)} 2, where f(x, p) is a regression, non-linear with respect to the parameter p, and satisfying some additional condition, given by a constraint r(f(x, p)) = 0. The problem is solved using a special recalculation of the weight function w(x) so that this constraint is taken into account. The method is the most appropriate and efficient for the solution of problems of decompositional filtering of data (function decomposition, robust regression analysis etc.).