R help - Nov 2003 - mixed model for Splus and R

Hi there,

I try to compare the mixed model package "lme" by Splus and R. I used
the
dataset "Ovary" and the following code assuming AR(1) model for the
error term:

lme(follicles ~ sin(2*pi*Time) + cos(2*pi*Time), data=Ovary, random = 
pdDiag(~sin(2*pi*Time) ) , correlation=corAR1() )

But I got different results! And then I used a simpler model:

lme(follicles ~ sin(2*pi*Time) + cos(2*pi*Time), data=Ovary, random = 
pdDiag(~sin(2*pi*Time)) )

This time I got the same output. I wonder the reason for the inconsistence 
between R and Splus. I attached the results as follows. Can someone tell me 
how to do it correctly in R? Thanks!

Lei

Result from R:

******************************************************************
Linear mixed-effects model fit by REML
   Data: Ovary
   Log-restricted-likelihood: -775.2224
   Fixed: follicles ~ sin(2 * pi * Time) + cos(2 * pi * Time)
        (Intercept) sin(2 * pi * Time) cos(2 * pi * Time)
         12.1895808         -2.9473432         -0.8807113

Random effects:
  Formula: ~sin(2 * pi * Time) | Mare
  Structure: Diagonal
         (Intercept) sin(2 * pi * Time) Residual
StdDev:    2.807293         0.03630784 3.665217

Correlation Structure: AR(1)
  Formula: ~1 | Mare
  Parameter estimate(s):
       Phi
0.6073908
Number of Observations: 308
Number of Groups: 11

*****************************************************************************************

Result from Splus:

*****************************************************************************************
Linear mixed-effects model fit by REML
   Data: Ovary
   Log-restricted-likelihood: -774.724
   Fixed: follicles ~ sin(2 * pi * Time) + cos(2 * pi * Time)
  (Intercept) sin(2 * pi * Time) cos(2 * pi * Time)
     12.18809          -2.985282         -0.8777629

Random effects:
  Formula:  ~ sin(2 * pi * Time) | Mare
  Structure: Diagonal
         (Intercept) sin(2 * pi * Time) Residual
StdDev:    2.858478            1.25802 3.507097

Correlation Structure: AR(1)
  Formula:  ~ 1 | Mare
  Parameter estimate(s):
        Phi
  0.5722019
Number of Observations: 308
Number of Groups: 11
 
******************************************************************************************


	[[alternative HTML version deleted]]

Hello

Anyone that now of any function in R that can calculate copulas?
Or if anyone have any code avaible I would be more than interested. 

Thank you in advance

/Thomas

Lei Liu <liulei at l.imap.itd.umich.edu> writes:
> I try to compare the mixed model package "lme" by Splus and R. I
used the
> dataset "Ovary" and the following code assuming AR(1) model for
the error term:
> 
> lme(follicles ~ sin(2*pi*Time) + cos(2*pi*Time), data=Ovary, random = 
> pdDiag(~sin(2*pi*Time) ) , correlation=corAR1() )
> 
> But I got different results! And then I used a simpler model:
> 
> lme(follicles ~ sin(2*pi*Time) + cos(2*pi*Time), data=Ovary, random = 
> pdDiag(~sin(2*pi*Time)) )
> 
> This time I got the same output. I wonder the reason for the inconsistence 
> between R and Splus. I attached the results as follows. Can someone tell me
> how to do it correctly in R? Thanks!
> 
> Lei
> 
> Result from R:
> 
> ******************************************************************
> Linear mixed-effects model fit by REML
>    Data: Ovary
>    Log-restricted-likelihood: -775.2224
>    Fixed: follicles ~ sin(2 * pi * Time) + cos(2 * pi * Time)
>         (Intercept) sin(2 * pi * Time) cos(2 * pi * Time)
>          12.1895808         -2.9473432         -0.8807113
> 
> Random effects:
>   Formula: ~sin(2 * pi * Time) | Mare
>   Structure: Diagonal
>          (Intercept) sin(2 * pi * Time) Residual
> StdDev:    2.807293         0.03630784 3.665217
> 
> Correlation Structure: AR(1)
>   Formula: ~1 | Mare
>   Parameter estimate(s):
>        Phi
> 0.6073908
> Number of Observations: 308
> Number of Groups: 11
> 
>
*****************************************************************************************
> 
> Result from Splus:
> 
>
*****************************************************************************************
> Linear mixed-effects model fit by REML
>    Data: Ovary
>    Log-restricted-likelihood: -774.724
>    Fixed: follicles ~ sin(2 * pi * Time) + cos(2 * pi * Time)
>   (Intercept) sin(2 * pi * Time) cos(2 * pi * Time)
>      12.18809          -2.985282         -0.8777629
> 
> Random effects:
>   Formula:  ~ sin(2 * pi * Time) | Mare
>   Structure: Diagonal
>          (Intercept) sin(2 * pi * Time) Residual
> StdDev:    2.858478            1.25802 3.507097
> 
> Correlation Structure: AR(1)
>   Formula:  ~ 1 | Mare
>   Parameter estimate(s):
>         Phi
>   0.5722019
> Number of Observations: 308
> Number of Groups: 11
>  
******************************************************************************************
S-PLUS and R use different optimizer code so the results will not be
identical.  You may want to check with intervals() to see how
well-defined the optima are.  Also try setting
control=list(msVerbose=TRUE) in the call to lme to see how the actual
optimization is behaving.

The log-likelihood in complex lme and nlme models can be a difficult
function to optimize.  I am aware that the results from R's optimizer(s)
are not always as good as those from the optimizer used in S-PLUS and
we are working on rectifying that problem.

On 18 Nov 2003 at 11:26, Thomas Holm wrote:

What is copulas?

Kjetil Halvorsen
> Hello
> 
> Anyone that now of any function in R that can calculate copulas?
> Or if anyone have any code avaible I would be more than interested. 
> 
> Thank you in advance
> 
> /Thomas
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help

Hi everybody,

Sorry for the OffTopic, but I always have a problem using apt-get to 
update my debian siystem and R. 
Anytime it updates the packages (right now I installed a self-compiled 
version of 1.8.1), even if they are exactly the same. Anybody can help 
me?

TIA,
Stefano

> Can you give us more details, please?  Which version of Debian
> (stable, testing or unstable) are you running and what are the
> relevant parts of your /etc/apt/sources.list file?  (Just send us the
> whole file if you are not sure what parts are relevant.)
Ok. Find attach my source.list + the output from apt-cache show 
r-base-core + apt-cache policy r-base-core.
I use Debian/testing version + some unstable package.
> 
> To others on the cc: list:  May I suggest that we ask Martin to create
> an email list with a name like r-debian for Debian-specific questions
> about R installation?
Yes, it would be a great idea!

Thanks,
Ste
> 
> -- 
> Douglas Bates                            bates at stat.wisc.edu
> Statistics Department                    608/262-2598
> University of Wisconsin - Madison        http://www.stat.wisc.edu/~bates/
-- 
Stefano Calza,
Sezione di Statistica Medica
Dip. di Scienze Biomediche e Biotecnologie
Universit? degli Studi di Brescia - Italy
Viale Europa, 11 25123 Brescia
email: calza at med.unibs.it
Telefono/Phone: +390303717532
Fax: +390303701157
-------------- next part --------------
deb http://ftp.it.debian.org/debian/ stable main non-free contrib
deb-src http://ftp.it.debian.org/debian/ stable main non-free contrib
deb http://non-us.debian.org/debian-non-US stable/non-US main contrib non-free
deb-src http://non-us.debian.org/debian-non-US stable/non-US main contrib
non-free

deb http://ftp.it.debian.org/debian/ testing main non-free contrib
deb http://ftp.it.debian.org/debian-non-US testing/non-US main contrib non-free
deb-src http://ftp.it.debian.org/debian testing main contrib non-free
deb-src http://ftp.it.debian.org/debian-non-US testing/non-US main contrib
non-free

deb http://ftp.it.debian.org/debian/ unstable main non-free contrib
deb http://ftp.it.debian.org/debian-non-US unstable/non-US main contrib non-free
deb-src http://ftp.it.debian.org/debian unstable main contrib non-free
deb-src http://ftp.it.debian.org/debian-non-US unstable/non-US main contrib
non-free

# R

#deb http://cran.at.r-project.org/bin/linux/debian stable main
#deb http://cran.at.r-project.org/bin/linux/debian testing main
#deb http://cran.at.r-project.org/bin/linux/debian unstable main

deb http://security.debian.org/ stable/updates main contrib non-free


## Java Environment

#deb http://mirrors.publicshout.org/java-linux/debian testing main non-free
#deb http://mirrors.publicshout.org/java-linux/debian unstable main non-free

## Bioconductor

deb http://lab.analytics.washington.edu/debian-local ./
-------------- next part --------------
Package: r-base-core
Priority: optional
Section: math
Installed-Size: 23916
Maintainer: Dirk Eddelbuettel <edd at debian.org>
Architecture: i386
Source: r-base
Version: 1.8.0.cvs.20031114-1
Replaces: r-base (<= 1.4.1-1)
Depends: perl, zlib-bin, libbz2-1.0, libc6 (>= 2.3.2.ds1-4), libg2c0 (>=
1:3.3.2-1), libgcc1 (>= 1:3.3.2-1), libjpeg62, libncurses5 (>=
5.3.20030510-1), libpcre3 (>= 4.0), libpng10-0 (>= 1.0.15-4), libreadline4
(>= 4.3-1), tcl8.4 (>= 8.4.2), tk8.4 (>= 8.4.2), xlibs (>>
4.1.0), zlib1g (>= 1:1.1.4)
Recommends: r-recommended, r-base-dev
Suggests: libpaperg, ess, r-doc-info | r-doc-pdf | r-doc-html
Filename: pool/main/r/r-base/r-base-core_1.8.0.cvs.20031114-1_i386.deb
Size: 5939210
MD5sum: 8cc59556c1385f4ec50b84f69380c691
Description: GNU R core of statistical computing language and environment
 R is `GNU S' - A language and environment for statistical computing
 and graphics. R is similar to the award-winning S system, which was
 developed at Bell Laboratories by John Chambers et al. It provides a
 wide variety of statistical and graphical techniques (linear and
 nonlinear modelling, statistical tests, time series analysis,
 classification, clustering, ...).
 .
 R is designed as a true computer language with control-flow
 constructions for iteration and alternation, and it allows users to
 add additional functionality by defining new functions. For
 computationally intensive tasks, C, C++ and Fortran code can be
 linked and called at run time.
 .
 S is the statistician's Matlab and R is to S what Octave is to Matlab.
 .
 This package provides the core GNU R system from which only the optional
 documentation packages r-base-html, r-base-latex, r-doc-html, r-doc-pdf
 and r-doc-info have been split off to somewhat reduce the size of this
 package.

Package: r-base-core
Status: install ok installed
Priority: optional
Section: math
Installed-Size: 23912
Maintainer: Dirk Eddelbuettel <edd at debian.org>
Source: r-base
Version: 1.8.0.cvs.20031114-1
Replaces: r-base (<= 1.4.1-1)
Depends: perl, zlib-bin, libbz2-1.0, libc6 (>= 2.3.2.ds1-4), libg2c0 (>=
1:3.3.2-1), libgcc1 (>= 1:3.3.2-1), libjpeg62, libncurses5 (>=
5.3.20030510-1), libpcre3 (>= 4.0), libpng12-0 (>= 1.2.5.0-4),
libreadline4 (>= 4.3-1), tcl8.4 (>= 8.4.2), tk8.4 (>= 8.4.2), xlibs
(>> 4.1.0), zlib1g (>= 1:1.1.4)
Recommends: r-recommended, r-base-dev
Suggests: libpaperg, ess, r-doc-info | r-doc-pdf | r-doc-html
Conffiles:
 /etc/R/Makeconf c819eaeb69d4f1514795e3a9eb7e2c9c
 /etc/R/Renviron 8ca0b400f24f00d60a2bacdcbb03db85
Description: GNU R core of statistical computing language and environment
 R is `GNU S' - A language and environment for statistical computing
 and graphics. R is similar to the award-winning S system, which was
 developed at Bell Laboratories by John Chambers et al. It provides a
 wide variety of statistical and graphical techniques (linear and
 nonlinear modelling, statistical tests, time series analysis,
 classification, clustering, ...).
 .
 R is designed as a true computer language with control-flow
 constructions for iteration and alternation, and it allows users to
 add additional functionality by defining new functions. For
 computationally intensive tasks, C, C++ and Fortran code can be
 linked and called at run time.
 .
 S is the statistician's Matlab and R is to S what Octave is to Matlab.
 .
 This package provides the core GNU R system from which only the optional
 documentation packages r-base-html, r-base-latex, r-doc-html, r-doc-pdf
 and r-doc-info have been split off to somewhat reduce the size of this
 package.

Package: r-base-core
Priority: optional
Section: math
Installed-Size: 21944
Maintainer: Dirk Eddelbuettel <edd at debian.org>
Architecture: i386
Source: r-base
Version: 1.7.1-1
Replaces: r-base (<= 1.4.1-1)
Depends: perl, zlib-bin, atlas2-base | atlas2, libbz2-1.0, libc6 (>=
2.3.1-1), libg2c0 (>= 1:3.3-0pre9), libgcc1 (>= 1:3.3-0pre9), libjpeg62,
libncurses5 (>= 5.3.20030510-1), libpcre3, libpng10-0 (>= 1.0.15-4),
libreadline4 (>= 4.3-1), tcl8.4 (>= 8.4.2), tk8.4 (>= 8.4.2), xlibs
(>> 4.1.0), zlib1g (>= 1:1.1.4)
Recommends: r-recommended, r-base-dev
Suggests: libpaperg, ess, r-doc-info | r-doc-pdf | r-doc-html
Filename: pool/main/r/r-base/r-base-core_1.7.1-1_i386.deb
Size: 5415844
MD5sum: c620e3533140e802038d0c0adce740a5
Description: GNU R core of statistical computing language and environment
 R is `GNU S' - A language and environment for statistical computing
 and graphics. R is similar to the award-winning S system, which was
 developed at Bell Laboratories by John Chambers et al. It provides a
 wide variety of statistical and graphical techniques (linear and
 nonlinear modelling, statistical tests, time series analysis,
 classification, clustering, ...).
 .
 R is designed as a true computer language with control-flow
 constructions for iteration and alternation, and it allows users to
 add additional functionality by defining new functions. For
 computationally intensive tasks, C, C++ and Fortran code can be
 linked and called at run time.
 .
 S is the statistician's Matlab and R is to S what Octave is to Matlab.
 .
 This package provides the core GNU R system from which only the optional
 documentation packages r-base-html, r-base-latex, r-doc-html, r-doc-pdf
 and r-doc-info have been split off to somewhat reduce the size of this
 package.

Package: r-base-core
Priority: optional
Section: math
Installed-Size: 19884
Maintainer: Dirk Eddelbuettel <edd at debian.org>
Architecture: i386
Source: r-base
Version: 1.5.1-0woody1
Replaces: r-base (<= 1.4.1-1)
Depends: perl, zlib-bin, atlas2-base | blas | blas2, libc6 (>= 2.2.4-4),
libjpeg62, libncurses5 (>= 5.2.20020112a-1), libpng2(>=1.0.12),
libreadline4 (>= 4.2a-4), tcl8.3 (>= 8.3.0), tk8.3 (>= 8.3.0), xlibs
(>> 4.1.0), zlib1g (>= 1:1.1.4)
Recommends: r-recommended, r-base-dev
Filename: pool/main/r/r-base/r-base-core_1.5.1-0woody1_i386.deb
Size: 4697006
MD5sum: ccfd24d8fc00c30c7965854b10516d63
Description: GNU R core of statistical computing language and environment
 R is `GNU S' - A language and environment for statistical computing
 and graphics. R is similar to the award-winning S system, which was
 developed at Bell Laboratories by John Chambers et al. It provides a
 wide variety of statistical and graphical techniques (linear and
 nonlinear modelling, statistical tests, time series analysis,
 classification, clustering, ...).
 .
 R is designed as a true computer language with control-flow
 constructions for iteration and alternation, and it allows users to
 add additional functionality by defining new functions. For
 computationally intensive tasks, C, C++ and Fortran code can be
 linked and called at run time.
 .
 S is the statistician's Matlab and R is to S what Octave is to Matlab.
 .
 This package provides the core GNU R system from which only the optional
 documentation packages r-base-html, r-base-latex, r-doc-html, r-doc-pdf
 and r-doc-info have been split off to somewhat reduce the size of this
 package.

-------------- next part --------------
r-base-core:
  Installato: 1.8.0.cvs.20031114-1
  Candidato: 1.8.0.cvs.20031114-1
  Tabella Versione:
     1.8.0.cvs.20031114-1 0
        500 http://ftp.it.debian.org unstable/main Packages
 *** 1.8.0.cvs.20031114-1 0
        100 /var/lib/dpkg/status
     1.7.1-1 0
        990 http://ftp.it.debian.org testing/main Packages
     1.5.1-0woody1 0
        500 http://ftp.it.debian.org stable/main Packages