R help - Aug 2005 - Help converting a function from S-Plus to R: family$weight

Hi all

I am converting an S-Plus function into R.  The S-Plus code
uses some of the glm families, and  family  objects.

The family objects in S-Plus and R have many different
features, for example:

In R:
 > names(Gamma())
  [1] "family"     "link"       "linkfun"   
"linkinv"    "variance"
  [6] "dev.resids" "aic"        "mu.eta"    
"initialize" "validmu"
[11] "valideta"

In S-Plus:
 > names(Gamma())
[1] "family"     "names"      "link"      
"inverse"    "deriv"
[6] "initialize" "variance"   "deviance"  
"weight"
 >


My question concerns the variable  weight  in the S-Plus function.
I'm not sure what it is.  (I have searched the S-Plus mailing list
archive, and my "S-Plus for linux 6.1" documentation.)  For almost all
family objects, the weight variable is the same as variance,
just weighted (and the former as a function; the later as an
expression):

 > Gamma()$variance
function(mu)
mu^2
 > Gamma()$weight
expression(w * mu^2.)
 >

The same applies for most families.  So I thought I could determine
what this weight variable was.

But alas--not the inverse,gaussian:

 > inverse.gaussian()$variance
function(mu)
mu^3
 > inverse.gaussian()$weight
expression(w/((sqrt(family$variance(mu)) * family$deriv(mu))^2.))


So:
- can anyone tell me what this expression weight represents?
- why is the inverse.gaussian family different than all others?

Thanks in advance.

P.

My S-Plus version:

 > version
Version 6.2.1  for Linux 2.4.18 : 2003

My R version:

 > version
          _
platform i386-pc-linux-gnu
arch     i386
os       linux-gnu
system   i386, linux-gnu
status
major    2
minor    1.0
year     2005
month    04
day      18
language R
 >


-- 
Dr Peter Dunn  |  Senior Lecturer in Statistics
Faculty of Sciences, University of Southern Queensland
   Web:    http://www.sci.usq.edu.au/staff/dunn
   Email:  dunn <at> usq.edu.au
CRICOS:  QLD 00244B |  NSW 02225M |  VIC 02387D |  WA 02521C

I think a question about the S family() functions is best determined by 
reading the S(-PLUS) documentation.  The weights component computes the 
working weights for the IWLS, unsurprisingly (and as stated by 
?family.object).  R has mu.eta to do that, the crucial line in R's glm.fit 
being

    w <- sqrt((weights[good] * mu.eta.val[good]^2)/variance(mu)[good])


But you don't need to port S-only features to R.  You do need to worry 
about R-only features.

On Fri, 12 Aug 2005, Peter Dunn wrote:
> Hi all
>
> I am converting an S-Plus function into R.  The S-Plus code
> uses some of the glm families, and  family  objects.
>
> The family objects in S-Plus and R have many different
> features, for example:
>
> In R:
> > names(Gamma())
>  [1] "family"     "link"       "linkfun"   
"linkinv"    "variance"
>  [6] "dev.resids" "aic"        "mu.eta"    
"initialize" "validmu"
> [11] "valideta"
>
> In S-Plus:
> > names(Gamma())
> [1] "family"     "names"      "link"      
"inverse"    "deriv"
> [6] "initialize" "variance"   "deviance"  
"weight"
> >
>
>
> My question concerns the variable  weight  in the S-Plus function.
> I'm not sure what it is.  (I have searched the S-Plus mailing list
> archive, and my "S-Plus for linux 6.1" documentation.)  For
almost all
> family objects, the weight variable is the same as variance,
> just weighted (and the former as a function; the later as an
> expression):
>
> > Gamma()$variance
> function(mu)
> mu^2
> > Gamma()$weight
> expression(w * mu^2.)
> >
>
> The same applies for most families.  So I thought I could determine
> what this weight variable was.
>
> But alas--not the inverse,gaussian:
>
> > inverse.gaussian()$variance
> function(mu)
> mu^3
> > inverse.gaussian()$weight
> expression(w/((sqrt(family$variance(mu)) * family$deriv(mu))^2.))
>
>
> So:
> - can anyone tell me what this expression weight represents?
> - why is the inverse.gaussian family different than all others?
>
> Thanks in advance.
>
> P.
>
> My S-Plus version:
>
> > version
> Version 6.2.1  for Linux 2.4.18 : 2003
>
> My R version:
>
> > version
>          _
> platform i386-pc-linux-gnu
> arch     i386
> os       linux-gnu
> system   i386, linux-gnu
> status
> major    2
> minor    1.0
> year     2005
> month    04
> day      18
> language R
> >
>
>
> -- 
> Dr Peter Dunn  |  Senior Lecturer in Statistics
> Faculty of Sciences, University of Southern Queensland
>   Web:    http://www.sci.usq.edu.au/staff/dunn
>   Email:  dunn <at> usq.edu.au
> CRICOS:  QLD 00244B |  NSW 02225M |  VIC 02387D |  WA 02521C
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
>
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595