[R] Iteratively Reweighted Least Squares of nonlinear regression

[Ravi Varadhan]

You are describing a "generalized nonlinear least-squares" estimation procedure.

This is implemented in the gnls() function in "nlme" package.

?gnls

Ravi.

____________________________________________________________________

Ravi Varadhan, Ph.D.
Assistant Professor,
Division of Geriatric Medicine and Gerontology
School of Medicine
Johns Hopkins University

Ph. (410) 502-2619
email: rvaradhan at jhmi.edu


----- Original Message -----
From: Derek An <derekan2 at gmail.com>
Date: Wednesday, July 1, 2009 11:50 am
Subject: [R] Iteratively Reweighted Least Squares of nonlinear regression
To: R-help at r-project.org


> Dear all,
>  
>  
>  When doing nonlinear regression, we normally use nls if e are iid normal.
>  
>    i learned that if the form of the variance of e is not completely known,
>  we can use the IRWLS (Iteratively Reweighted Least Squares )
>  
>  algorithm:
>  
>  for example, var e*i =*g0+g1*x*1
>  
>  1. Start with *w**i = *1
>  
>  2. Use least squares to estimate b.
>  
>  3. Use the residuals to estimate g, perhaps by regressing e^2 on *x*.
>  
>  4. Recompute the weights and goto 2.
>  
>  Continue until convergence
>  
>  i was wondering whether there is a instruction of R to do this?
>  
>  	[[alternative HTML version deleted]]
>  
>  ______________________________________________
>  R-help at r-project.org mailing list
>  
>  PLEASE do read the posting guide 
>  and provide commented, minimal, self-contained, reproducible code.