s-news
[Top] [All Lists]

Re: Hmisc multiple imputation with owner glm family

To: GIORGI Roch <roch.giorgi@mail.ap-hm.fr>
Subject: Re: Hmisc multiple imputation with owner glm family
From: Frank E Harrell Jr <f.harrell@vanderbilt.edu>
Date: Fri, 01 Jun 2007 07:24:43 -0500
Cc: "'s-news@lists.biostat.wustl.edu'" <s-news@lists.biostat.wustl.edu>
In-reply-to: <56A80768E1273C42B9C60D53285D501501A15D75@phoceo07.aphm.ap-hm.fr>
References: <56A80768E1273C42B9C60D53285D501501A15D75@phoceo07.aphm.ap-hm.fr>
User-agent: Thunderbird 1.5.0.10 (X11/20070403)
GIORGI Roch wrote:
Professor Harrel,

I've sent this email to the SPLUS list and I had no answer. Any help will be 
appreciate.

Sincerely,

Roch

The pertinent code inside fit.mult.impute is

    f <- if(missing(subset)) fitter(formula, data=completed.data, ...)
         else fitter(formula, data=completed.data[subset,], ...)

The family argument should be part of the ...

So I don't understand why glm is not seeing family. Make sure that your special family argument works correctly in calling glm outside of fit.mult.impute.

Frank

--------------------


I'm using Hmisc package to do multiple imputation.
I have to use glm method but with a family that is not yet part of S-PLUS. I created my own family but I don't know how to call it when I use Hmisc's fit.mult.impute.
I tried this:
mi <- aregImpute(~Y + sex + age1 + age2 + X1, n.impute=5, data=MCAR)
fit <- fit.mult.impute(Y ~ sex + age1 + age2 + X1, family="MyFamily", 
fitter=glm, xtrans=mi, data=MCAR)

but I have the impression that the family used is not MyFamily
fit$family gives:
name link variance "Gaussian" "Identity: mu" "Constant: 1"

I am using Splus6.2 and Hmisc 3.3-2.



_____________________________________________________
  Roch GIORGI, MD, PhD
  Service de Santé Publique et d'Information Médicale
  Hôpital de la Timone
264, rue St Pierre 13385 Marseille cedex 5 Tel: +33 (0)491 384 949 Fax: +33 (0)491 385 749 email: roch.giorgi@ap-hm.fr Web: http://cybertim.timone.univ-mrs.fr/




--
Frank E Harrell Jr   Professor and Chair           School of Medicine
                     Department of Biostatistics   Vanderbilt University

<Prev in Thread] Current Thread [Next in Thread>
  • Re: Hmisc multiple imputation with owner glm family, Frank E Harrell Jr <=