# nlmrt-package: Tools for solving nonlinear least squares problems.... In nlmrt: Functions for Nonlinear Least Squares Solutions (Deprecated!)

## Description

The package provides some tools related to using the Nash variant of Marquardt's algorithm for nonlinear least squares. This package has been replaced with package nlsr that has similar and improved capabilities, but with some differences in structure and syntax.

## Details

 Package: nlmrt Type: Package Version: 1.0 Date: 2012-03-05 License: GPL-2

This package includes methods for solving nonlinear least squares problems specified by a modeling expression and given a starting vector of named paramters. Note: You must provide an expression of the form lhs ~ rhsexpression so that the residual expression rhsexpression - lhs can be computed. The expression can be enclosed in quotes, and this seems to give fewer difficulties with R. Data variables must already be defined, either within the parent environment or else in the dot-arguments. Other symbolic elements in the modeling expression must be standard functions or else parameters that are named in the start vector.

The main functions in nlmrt are:

 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 nlfb - Nash variant of the Marquardt procedure for nonlinear least squares, with bounds constraints, using a residual and optionally Jacobian described as \code{R} functions. 20120803: Todo: Make masks more consistent between nlfb and nlxb. nlxb - Nash variant of the Marquardt procedure for nonlinear least squares, with bounds constraints, using an expression to describe the residual via an \code{R} modeling expression. The Jacobian is computed via symbolic differentiation. wrapnls - Uses nlxb to solve nonlinear least squares then calls nls() to create an object of type nls. model2grfun.R - Generate a gradient vector function from a nonlinear model expression and a vector of named parameters. model2jacfun.R - Generate a Jacobian matrix function from a nonlinear model expression and a vector of named parameters. model2resfun.R - Generate a residual vector function from a nonlinear model expression and a vector of named parameters. model2ssfun.R - Generate a sum of squares objective function from a nonlinear model expression and a vector of named parameters. modgr.R - compute gradient of the sum of squares function using the Jacobian and residuals for a nonlinear least squares problem modss.R - computer the sum of squares function from the residuals of a nonlinear least squares problem myfn.R, mygr.R, myjac.R, myres.R, myss.R - dummy functions that seem to be needed so there is an available handle for output of functions that generate various functions from expressions.

For testing purposes, there are also some experimental codes using different internal computations for the linear algebraic sub-problems in the inst/dev-codes/ sub-folder.

## Author(s)

John C Nash

Maintainer: <[email protected]>

## References

Nash, J. C. (1979, 1990) _Compact Numerical Methods for Computers. Linear Algebra and Function Minimisation._ Adam Hilger./Institute of Physics Publications

others!!??