| Author(s): I. Silin | Library: MATHLIB |
| Submitter: A. Kobine | Submitted: 05.04.1971 |
| Language: Fortran | Revised: 18.11.1985 |
and ML defined by:
and
with respect to the M parameters A where, for each j,
is a data-point with user estimated error,
, the 
are L co-ordinates of that point and Y
is a theoretical function predicting 
for a given set of 
and A.
The method makes use of a particular property concerning the
dependence of the objective function ( 
or ML) on the
theoretical function (Y) for faster convergence.
Structure:
SUBROUTINE subprograms
User Entry Names: FUMILI, LIKELM, ERRORF
Internal Entry Names: ARITHM, D510BD, FUNCT,
MCONV, MONITO, SCAL, SGZ
Files Referenced: Printer
External References: User-supplied FUNCTION and
(optional) SUBROUTINE subprograms
COMMON Block Names and Lengths: 
Usage:
See Long Write-up.
References:
