[R] "xtable" results doesn't correspond to data.frame
Dear useRs,
Am trying to use xtable on the following data.frame and I don't get what I
expect:
example.table <- data.frame(rbind(
c("Gender"," "," "," "),
cbind(rep(" ",2),c("Male","Female"),c(3.0,4.0),c(3/7,4/7))
))
colnames(example.table) <- c(" "," ","number of patients","%")
example.xtable <- xtable(example.table)
print.xtable(example.xtable, include.rownames=FALSE)
I can seem to get latex output which corresponds to the data.frame which is,
\begin{table}[ht]
\begin{center}
\begin{tabular}{}
\hline
& & number of patients & \% \\
\hline
Gender & & & \\
& Male & 3 & 0.428571428571429 \\
& Female & 4 & 0.571428571428571 \\
\hline
\end{tabular}
\end{center}
\end{table}
__
R-help@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
and provide commented, minimal, self-contained, reproducible code.
[R] problems with index within a strata
dear useRs,
I have several datasets which are in repeated measurements format and I
would like to extract the first measurement for each individual. In some
cases measurements of certain characteristics are also stacked on top of
each other and thats where the problem lies. A prototype of the code I use
is as follows,
# If I am extracting the first element for each individual (works fine)
temp <- sort(rpois(100,2.5))
temp.tab <- as.vector(table(temp))
cbind(temp,unlist(lapply(1:length(temp.tab), function(i){1:temp.tab[i]})))
# If I am extracting the first element of a variable for each individual
(doesn't work)
temp.0 <- sort(rpois(100,2.5))
temp.1 <- rbinom(100,1,0.25)
temp.tab <- as.vector(t(table(temp.0,temp.1)))
cbind(temp,unlist(lapply(1:length(temp.tab), function(i){1:temp.tab[i]})))
I don't understand why the first works ok and the second fails.
"sum(temp.tab)" is still the same.
Thanks for your help,
Vumani
__
R-help@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
and provide commented, minimal, self-contained, reproducible code.
[R] sorting elements of matrix by row
Dear R users,
I would like to sort elements of a matrix by row and use this ordering to
also sort another matrix. I am trying to post-order the means of components
for a mixture model and would also like to do the same for the component
probabilities. This is what I have tried thus far, but I doubt whether its
efficient given the number of replications I am using:
mean.data <- matrix(rnorm(20,0,1),10,2)
probs <- runif(10)
prob.data <- cbind(probs,1-probs)
sorted.mean <- sorted.prob <- array(0, dim=dim(mean.data))
for(i in 1:nrow(mean.data)){
sorted.data[i,] <- mean.data[i,order(mean.data[i,])]
sorted.prob[i,] <- prob.data[i,order(mean.data[i,])]
}
Thanks,
Vumani
__
R-help@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
and provide commented, minimal, self-contained, reproducible code.
[R] i suspect that there a memory leak in "vmmin"?
Dear listers, Am currently using MCMC approaches to estimate some parameters of my model. One parameter has to be updated using a tuned gamma distribution. So at each iteration I estimate the mean and variance of the density of the gamma approximation using "vmmin" (i also supply the gradient argument). For moderate replications the procedure works, but if I increase them R crashes. If instead of the tuned gamma density i use the adaptive rejection sampling procedure by Gilks and Wild (downloaded from their website) the procedure returns the results for any number of replications, thus I suspected that there was a memory leak in "vmmin". I am not an expert in "C" and am only a windows user. Have tried some of the debugging tools available to us windows users: gdb, mpatrol, duma without success. Am not sure whether this is a bug. Version 2.3.1 (2006-06-01) Windows XP Regards, Vumani Rgui.exe caused an Access Violation at location 100dec2f in module R.dll Reading from location 3ff7c27d. Registers: eax=63ac ebx=01e773a0 ecx=3ff7c27a edx=0002 esi=01e254b4 edi=0002 eip=100dec2f esp=00e0ecb0 ebp=00e0ed78 iopl=0 nv up ei pl nz na pe nc cs=001b ss=0023 ds=0023 es=0023 fs=003b gs= efl=0202 Call stack: 100DEC2F R.dll:100DEC2F SET_TAG 100E0527 R.dll:100E0527 Rf_allocVector 100E0D37 R.dll:100E0D37 R_alloc 100F2ACA R.dll:100F2ACA vmmin 023F5E96 gem.dll:023F5E96 weibullMH 10088C33 R.dll:10088C33 do_dotcall 100B8827 R.dll:100B8827 Rf_eval 100BA4D2 R.dll:100BA4D2 do_set 100B86CB R.dll:100B86CB Rf_eval 100BA5D5 R.dll:100BA5D5 do_begin 100B86CB R.dll:100B86CB Rf_eval 100BB8EB R.dll:100BB8EB Rf_applyClosure 100B85F8 R.dll:100B85F8 Rf_eval 100BA4D2 R.dll:100BA4D2 do_set 100B86CB R.dll:100B86CB Rf_eval 100DB51C R.dll:100DB51C Rf_ReplIteration 100DBAA6 R.dll:100DBAA6 run_Rmainloop 004013CF Rgui.exe:004013CF 00401316 Rgui.exe:00401316 00401518 Rgui.exe:00401518 00401236 Rgui.exe:00401236 00401288 Rgui.exe:00401288 7C816D4F kernel32.dll:7C816D4F RegisterWaitForInputIdle __ R-help@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
[R] using "rbinom" in C code gives me erroneous results... random variable is not random (always zero)...
Dear Listers,
I am trying to use "rbinom" in my C code, but i always get zeros as output no
matter the probability. Am not sure what I am doing wrong because the function
has worked before. Attached in an example. Noticed that "rbinom" expects 'n' to
be REAL.
Regards, Vumani
R 2.3.1 (2006-06-01)
Windows XP
Gcc
/* Called this file binom.c and then ran rcmd shlib on it */#include
#include #include #include SEXP binomial(SEXP
r, SEXP n, SEXP p){ int i; SEXP out; PROTECT(out =
allocVector(REALSXP,INTEGER(r)[0])); for(i = 0; i < INTEGER(r)[0]; ++i){
REAL(out)[i] = rbinom(REAL(n)[0],REAL(p)[0]); } Rprintf("%d %f
%f\n",INTEGER(r)[0],REAL(n)[0],REAL(p)[0]); UNPROTECT(1); return(out);}
### used theses line in R
dyn.load("binom.dll").Call("binomial",as.integer(10),as.double(1),as.double(0.4))dyn.unload("binom.dll")
_
Try Live.com - your fast, personalized homepage with all the things you care
about in one place.
http://www.live.com/getstarted
[[alternative HTML version deleted]]
__
R-help@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
[R] fitting a distribution using glm
it is easy to fit a distribution using fitdistr
poisdata <- rpois(n = 100, lambda = 2)
poismle <- fitdistr(poisdata, "Poisson")
poismle
but i would like to know whether its possible to get an identical result
using glm. I use
poistab <- data.frame(table(poisdata))
colnames(poistab) <- c("width","freq");
poistab[,"width"] <- as.numeric(poistab[,"width"])
glm(freq ~ 1-width, family = poisson, data = poistab)
but the results are always different.
cheers, vumani
__
R-help@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
[R] R function use in C returning a list
Dear R users,
Is it possible to use an R function accepting several arguments and
returning a list in C code. I notice that in all examples only one variable
is returned.
Here is the C part I feel is giving the problems,
SEXP R_fcall, ans;
PROTECT(R_fcall = lang2(likelihood, R_NilValue));
++nProtected;
defineVar(install("PAR"), PAR, env);
defineVar(install("DATA"), oldData, env);
defineVar(install("N"), N, env);
PROTECT(ans = eval(R_fcall, env));
oldData is a matrix which i set using
SEXP dim;
PROTECT(dim = NEW_INTEGER(2));
++nProtected;
INTEGER_DATA(dim)[0] = n;
INTEGER_DATA(dim)[1] = 2;
setAttrib(oldData, R_DimSymbol, dim);
if i return both function and data they are ok, but its evaluating the
function and returning the list or even a single number which is a problem.
Thanks again, Vumani
__
R-help@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
[R] RCMD SHILB uses g++ instead of gcc
I am using RCMD SHLIB to make a dll from the convolve2.c code presented with
R-exts. After an e-mail from Prof Ripley, I noticed that SHLIB is using a
C++ compiler instead of a C compiler for the code.
Have searched the R-site and the internet for answers and noticed that it is
possible to change the "makeconf" and "makefile" files to correct this. But
these files are not present in my system. I downloaded
"R-2.2.1pat-win32.exe" and installed the traditional windows way.
Thanks for your responses.
Full_Name: Vumani Dlamini
Version: 2.2.1 Patched
OS: Microsoft Windows XP Professional
>From: Prof Brian Ripley <[EMAIL PROTECTED]>
>To: Vumani Dlamini <[EMAIL PROTECTED]>
>CC: r-help@stat.math.ethz.ch
>Subject: Re: [R] symbol decoration and .Call
>Date: Thu, 16 Feb 2006 13:12:25 + (GMT)
>
>1) Please read the posting guide and use the appropriate list, not R-help.
>
>2) Please tell us exactly what you did, including your OS, R version and so
>on as the guide asks. Looks to me like you used a C++ compiler to compile
>C.
>
>On Thu, 16 Feb 2006, Vumani Dlamini wrote:
>
>>Dear R-users,
>>I am a novice user of .Call and am trying to use the C code in R-Ext to
>>kickstart my learning process. All the code compiles but when I try to use
>>.Call it gives the error (using "out.c" as an example),
>>
>>Error in .Call("out", x, y) : C entry point "out" not in load table
>>
>>when i use PEDUMP to check whether the symbol is loaded I find that it is
>>loaded as "_Z3outP7SEXPRECS0_" and when I use this instead of "out" the
>>program work.
>>
>>How can I remove the decoration so that only "out" is loaded?
>>
>>Thanks, Vumani
>>
>>__
>>R-help@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, [EMAIL PROTECTED]
>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, UKFax: +44 1865 272595
__
R-help@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
[R] symbol decoration and .Call
Dear R-users,
I am a novice user of .Call and am trying to use the C code in R-Ext to
kickstart my learning process. All the code compiles but when I try to use
.Call it gives the error (using "out.c" as an example),
Error in .Call("out", x, y) : C entry point "out" not in load table
when i use PEDUMP to check whether the symbol is loaded I find that it is
loaded as "_Z3outP7SEXPRECS0_" and when I use this instead of "out" the
program work.
How can I remove the decoration so that only "out" is loaded?
Thanks, Vumani
__
R-help@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
[R] finding roots of a polynomial
Dear listers, I am trying to find the root of a polynomial of the form, a+bx+cx^2+dx^3+ex^5+...=0 where a,b,c,d,e are known. I have looked at "quadprog" without any success. Regards, Vumani __ R-help@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
[R] drawing a network digraph
Dear R users,
I have a matrix with 2 columns with the variables: "daughter index", "mother
index". I would like to draw a network digraph using this data, where each
daughter is connected to a mother and between the connections inlcude a
circle with the information on the indices ("daughter index", "mother
index"): i.e. something similar to graphs produced by graphviz.
I am clueless. I have looked at the libraries SEM and SNA but it looks like
they can't help my cause.
Hope you help.
Vumani
__
R-help@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
[R] "\n" in legend using substitute and as.expression
Dear R users,
Is it possible to force a hard return in an expression. I tried including
one for a legend using an expression and it didn't work. Here is my code,
legend(12, 0.10, c("Fitted density",
as.expression(substitute(paste("Weibull (", alpha==shapeU," , ",
beta==scaleU,") Mean = ",meanU," days"),
list(shapeU = round(mle.r$estimate[1],3),
scaleU = round(mle.r$estimate[2],3),
meanU = round(mle.r$estimate[3),4 ),
lty=c(1,1),col=c(1,2),lwd=c(1,1),merge = TRUE)
(Windows XP + R2.10 user)
Thanks,
Vumani
__
R-help@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
[R] including R function in C program
Is it possoble to include an R function in a C program? I would like to use "rmultz" from the "combinat" library to sample a multinomial vector within my C program, and have no clue how this can be done. Thanks. Vumani __ R-help@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
[R] DLL problem when using gcc+gsl
Dear list;
This might sound a bit naive, but then I am new to linking C DLL's to R.
I have built a DLL using GCC and am able to load the DLL in R
(is.loaded("contents")==TRUE). When I include it in my function it returns
NaN for all the variables. My R function is,
dyn.load("c:/data/tempdll.dll")
is.loaded("contents")#returns TRUE
#
contents <- function(new.cases){
cases <- rep(0, length(new.cases))
case.times <- rep(0, length(new.cases))
temp <- .C("contents",
as.integer(cases),
as.double(case.times))
}
and the header of my C function is
#ifndef _DLL_H_
#define _DLL_H_
#if BUILDING_DLL
# define DLLIMPORT __declspec (dllexport)
#else /* Not BUILDING_DLL */
# define DLLIMPORT __declspec (dllimport)
#endif /* Not BUILDING_DLL */
DLLIMPORT void contents (int *cases,
double *case_times);
#endif /* _DLL_H_ */
and my C function is.
#include
#include
#include
#include "dll.h"
#include
#include
void contents(int *cases,
double *case_times)
{
gsl_rng * r;
gsl_rng_env_setup();
r = gsl_rng_alloc (gsl_rng_default);
int i;
for(i = 0; i < 1000; i++)
{
cases[i] = gsl_ran_negative_binomial (r, 0.10/2.10, 0.10);
case_times[i] = gsl_ran_weibull (r, 9, 5);
}
gsl_rng_free (r);
}
I am clueless as to where it all goes wrong, as if I make the program an
executable it works.
Thanks again.
Vumani
__
R-help@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
[R] sampling from a mixture distribution
Dear R users, I would like to sample from a mixture distribution p1*f(x1)+p2*f(x2). I usually sample variates from both distributions and weight them with their respective probabilities, but someone told me that was wrong. What is the correct way? Vumani __ R-help@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
[R] counting number of objects in a list
Dear R-users; I am interested in getting the number of objects in a list, and have thus far been unsuccessful. Can you please help. Thank you. Vumani __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] manipulating elements of a vector
Dear R users; I would like to convert a series of vectors to matrices in the following way; (2,1,1) to a matrix 1 0 0 1 0 0 0 1 0 0 0 1 The idea is that the column sum of the matrix should be equal to the elements of the vector. Thanks. Vumani __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] subsetting based on vector
Thanks man. It did the trick.
Vumani
From: Winfried Theis <[EMAIL PROTECTED]>
To: Vumani Dlamini <[EMAIL PROTECTED]>
CC: [EMAIL PROTECTED]
Subject: Re: [R] subsetting based on vector
Date: Thu, 25 Mar 2004 16:16:58 +0100
Hi!
Well, you could try a subsetting rule like " year %in% yearsconsidered
".
Hope this helps,
Winfried
On Thu, 2004-03-25 at 14:33, Vumani Dlamini wrote:
> Dear R users;
>
> I am trying to write a small program which reads in a data set, and
selects
> observations from certain years before the analysis. I have a problem
> including the selection criteria in the header of the program.
>
> Here is the problem;
>
> dataFIT<-function(MODEL, MARGINS, yearConsidered){
> library(foreign
> CovPaper<-read.spss("C:/Data/data.sav")
> NewData <- list(CovPaper$"YEAR"[CovPaper$"YEAR" == yearConsidered],
> CovPaper$"YEAR"[CovPaper$"SEX" ==
yearConsidered],
> and so on #
>#fit model to data #
> }
>
> When I use one year there is no problem, but I would like some data sets
to
> span over years and I am not sure how to do this without having to
change
> the body of my fitting program.
>
> I tried searching the R-list but to no avail.
>
> Thanking you in advance.
>
>
> Vumani
>
> _
> Hotmail en Messenger on the move
> http://www.msn.nl/communicatie/smsdiensten/hotmailsmsv2/
>
> __
> [EMAIL PROTECTED] mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
--
-
Dr. Dipl.-Math. Winfried Theis, SFB 475, Projekt C5,
Universität Dortmund, 44221 Dortmund
e-mail: [EMAIL PROTECTED]
Tel.: +49/231/755-5903 FAX: +49/231/755-4387
__
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] subsetting based on vector
Dear R users;
I am trying to write a small program which reads in a data set, and selects
observations from certain years before the analysis. I have a problem
including the selection criteria in the header of the program.
Here is the problem;
dataFIT<-function(MODEL, MARGINS, yearConsidered){
library(foreign
CovPaper<-read.spss("C:/Data/data.sav")
NewData <- list(CovPaper$"YEAR"[CovPaper$"YEAR" == yearConsidered],
CovPaper$"YEAR"[CovPaper$"SEX" == yearConsidered],
and so on #
#fit model to data #
}
When I use one year there is no problem, but I would like some data sets to
span over years and I am not sure how to do this without having to change
the body of my fitting program.
I tried searching the R-list but to no avail.
Thanking you in advance.
Vumani
_
Hotmail en Messenger on the move
http://www.msn.nl/communicatie/smsdiensten/hotmailsmsv2/
__
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] (no subject)
Dear R users, I am trying to convert a dataset from one format to several rectangular datasets. A consultant helped design the data entry program for our survey using Delphi/Pascal and for each household the information is stored in a file called "EA-HM-HH.TXT" where EA is the enumeration area number, HM is the homestead number and HH is the household number. Within this file the data is stored as follows, ## file="5677-001-001.TXT" ## EAnumber=5677 HMnumber=001 HHnumber=001 # Demographics section Dserial=01 #first person in household Dage=56 Dsex=1 Dserial=02 #second person in household Dage=44 Dsex=2 Dserial=03 #second person in household Dage=7 Dsex=2 # Agricultural inputs section Amaize=200 Apumpkins=50 ## end of file Note that in the demograpics section there may only be less than 3 or more people in some households. I would like to create a file for the demographics section which is as follows EAnumber|HMnumber|HHnumber|Dserial|Dage|Dsex 5677 001 001 056 1 5677 001 001 044 2 5677 001 001 007 2 and for the agricultural inputs I would like to have, EAnumber|HMnumber|HHnumber|Amaize|Apumpkins 5677 001 001 0200 5677 001 001 0050 There are several similar files where the EA number, HM number or HH number changes, thus I would also like to know whether it is possible to create a script where for all household there is only one dataset for demographics, and one for agricultural inputs. Thanking you as always. Vumani Dlamini __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] query on converting survey data from one structure to another
Dear R users, I am trying to convert a dataset from one format to several rectangular datasets. A consultant helped design the data entry program for our survey using Delphi/Pascal and for each household the information is stored in a file called "EA-HM-HH.TXT" where EA is the enumeration area number, HM is the homestead number and HH is the household number. Within this file the data is stored as follows, ## file="5677-001-001.TXT" ## EAnumber=5677 HMnumber=001 HHnumber=001 # Demographics section Dserial=01 #first person in household Dage=56 Dsex=1 Dserial=02 #second person in household Dage=44 Dsex=2 Dserial=03 #second person in household Dage=7 Dsex=2 # Agricultural inputs section Amaize=200 Apumpkins=50 ## end of file Note that in the demograpics section there may only be less than 3 or more people in some households. I would like to create a file for the demographics section which is as follows EAnumber|HMnumber|HHnumber|Dserial|Dage|Dsex 5677 001 001 056 1 5677 001 001 044 2 5677 001 001 007 2 and for the agricultural inputs I would like to have, EAnumber|HMnumber|HHnumber|Amaize|Apumpkins 5677 001 001 0200 5677 001 001 0050 There are several similar files where the EA number, HM number or HH number changes, thus I would also like to know whether it is possible to create a script where for all household there is only one dataset for demographics, and one for agricultural inputs. Thanking you as always. Vumani Dlamini __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Hmisc multiple imputation functions
Dear all;
I am trying to use HMISC imputation function to perform multiple imputations
on my data and I keep on getting errors for the code given in the help
files.
When using "aregImpute" the error is;
f <- aregImpute(~y + x1 + x2 + x3, n.impute=100)
Loading required package: acepack
Iteration:1 Error in .Fortran("wclosepw", as.double(w), as.double(x),
as.double(runif(lw)), :
C/Fortran function name not in load table
When I use "transcan" it is;
f <- transcan(~y + x1 + x2, n.impute=10, shrink=FALSE)
Error in transcan(~y + x1 + x2, n.impute = 10, shrink = FALSE) :
Must specify data= when using R
I am not sure what I am missing.
Vumani
__
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] memory problems in NLME
Dear R-Users I am trying to fit a random coeficient model with about 30 covariates with a random intercept and one random slope. The data set has 65000 observations, and whenever I use LME I get the message that all memory has been used. I was wondering whether there is a more efficient way fo fitting the model. OS: Windows 2000 R version: R 1.6.2 Memory: 128Mb Free HD: <>11Gb Looking forward to you help as always. Vumani Dlamini __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] random error and grouping variable correlated
Dear R-users: Is it possible to fit linear mixed models using LME when the random error is correlated with one of the covariates in the model? Thanks. Vumani Dlamini __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Simple NLME problem (I hope)
Dear R-Users: I would like to fit a multilevel model using LME such that the parameters after fitting the multilevel in two separate groups and when I use the complete data (with interactions between the grouping variable and the other variables) set are comparable (or the same). The problem I am having currently is that I am not sure whether it is possible to let the random error term to vary by group such that the models are comparable. At present only one random error term is estimate and this "sort of" distorts the parameter estimates for the fixed and other random effects. Looking forward to your help. Vumani Dlamini __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] expression for simple EM
Dear R users, I know this is the wrong forum for such a question but I need help. I would like to write down the likelihood expression for a simple EM problem. I have one categorical covariate with 5 levels and a missing count which can fall in any of the categories. I know the solution to the problem but can't seem to get the likelihood expression correct (if it is possible). Thanks. Vumani Dlamini, Student Faculty of Health Sciences University of Natal __ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] something wrong when using pspline in clogit?
Dear R users: I am not entirely convinced that clogit gives me the correct result when I use pspline() and maybe you could help correct me here. When I add a constant to my covariate I expect only the intercept to change, but not the coefficients. This is true (in clogit) when I assume a linear in the logit model, but the same does not happen when I use pspline(). If I did something similar to what "na.gam.replace" in S+ does, the pspline coefficients are the same, leading me to believe that "na.gam.replace" is doing the right thing (I may be wrong). The strange thing here is that even the intercepts are the same. Lastly, these coefficients are the same as using clogit before adding the constant, but the intercept are different. I have attached my code for all this, maybe I messed it somewhere. Vumani ## xvar<-rnorm(100,0,1) data.mult<-rbinom(100,size=3,prob=(exp(0.5+0.4*xvar)/(1+exp(0.5+0.4*xvar library(Nnet) mult.fit<-multinom(data.mult~xvar) coef(mult.fit) library(survival) choice<-c(ifelse(data.mult==0,1,0),ifelse(data.mult==1,1,0),ifelse(data.mult==2,1,0),ifelse(data.mult==3,1,0)) temp<-list(time=2-choice, status=choice, intercept1=c(rep(0,length(data.mult)),rep(1,length(data.mult)),rep(0,length(data.mult)),rep(0,length(data.mult))), intercept2=c(rep(0,length(data.mult)),rep(0,length(data.mult)),rep(1,length(data.mult)),rep(0,length(data.mult))), intercept3=c(rep(0,length(data.mult)),rep(0,length(data.mult)),rep(0,length(data.mult)),rep(1,length(data.mult))), xvar1=c(rep(0,length(data.mult)),rep(1,length(data.mult)),rep(0,length(data.mult)),rep(0,length(data.mult)))*rep(xvar,4), xvar2=c(rep(0,length(data.mult)),rep(0,length(data.mult)),rep(1,length(data.mult)),rep(0,length(data.mult)))*rep(xvar,4), xvar3=c(rep(0,length(data.mult)),rep(0,length(data.mult)),rep(0,length(data.mult)),rep(1,length(data.mult)))*rep(xvar,4), stratum=rep(1:length(data.mult),4)) # Used to check whether my data matrix is setup correctly # This fits a category one baseline category logit model # Compare with multinom totalmodel<-clogit(status~intercept1+intercept2+intercept3+xvar1+xvar2+xvar3+strata(stratum),method="exact",temp) coef(totalmodel) # Use smooth term, "pspline" totalmodel<-clogit(status~intercept1+intercept2+intercept3+pspline(xvar1)+pspline(xvar2)+pspline(xvar3)+strata(stratum),method="exact",temp) coef(totalmodel) # Collect indices where variable is not equal to zero indices<-list(var1=NULL,var2=NULL,var3=NULL) indices$var1<-temp$xvar1==0 indices$var2<-temp$xvar2==0 indices$var3<-temp$xvar3==0 # Add 10 to covariate temp$xvar1[!indices$var1]<-xvar+10 temp$xvar2[!indices$var2]<-xvar+10 temp$xvar3[!indices$var3]<-xvar+10 # First fit linear-in-the-logit model, to check what happens to coefficients totalmodel<-clogit(status~intercept1+intercept2+intercept3+xvar1+xvar2+xvar3+strata(stratum),method="exact",temp) coef(totalmodel) # Fit a smooth model, and compare with smooth model above totalmodel<-clogit(status~intercept1+intercept2+intercept3+pspline(xvar1)+pspline(xvar2)+pspline(xvar3)+strata(stratum),method="exact",temp) coef(totalmodel) # Doing something similar to what na.gam.replace() in S+ does temp$xvar1[indices$var1]<-mean(xvar+10) temp$xvar2[indices$var2]<-mean(xvar+10) temp$xvar3[indices$var3]<-mean(xvar+10) # Fit spline model totalmodel<-clogit(status~intercept1+intercept2+intercept3+pspline(xvar1)+pspline(xvar2)+pspline(xvar3)+strata(stratum),method="exact",temp) coef(totalmodel) # Remove 10 and substitute with mean # Covariate are like in the beginning temp$xvar1[!indices$var1]<-xvar temp$xvar2[!indices$var2]<-xvar temp$xvar3[!indices$var3]<-xvar # Doing something like na.gam.replace() temp$xvar1[indices$var1]<-mean(xvar) temp$xvar2[indices$var2]<-mean(xvar) temp$xvar3[indices$var3]<-mean(xvar) totalmodel<-clogit(status~intercept1+intercept2+intercept3+pspline(xvar1)+pspline(xvar2)+pspline(xvar3)+strata(stratum),method="exact",temp) coef(totalmodel) From: [EMAIL PROTECTED] To: Vumani Dlamini <[EMAIL PROTECTED]> CC: [EMAIL PROTECTED] Subject: Re: [R] R analogue Date: Mon, 20 Jan 2003 13:17:02 + (GMT) On Mon, 20 Jan 2003, Vumani Dlamini wrote: > Is there any R analogue for the S+ function "na.gam.replace". No, for it is tailored for use by S's gam. Some of the things it does are positively undesirable! It uses mean imputation for continuous variables, but for factors it makes NA into another level, which silently assumes that all missing values are similar and that they are going to occur with sufficient frequency in the training data (and they may not occur at all). > I would like > to make an interaction of a categorical and smooth continuous covariate.
[R] R analogue
Dear R users: Is there any R analogue for the S+ function "na.gam.replace". I would like to make an interaction of a categorical and smooth continuous covariate. Thanks. Vumani Dlamini Central Statistical Office Swaziland __ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Re: Universal legend in plot
Dear R-users: I asked a question on how I can have a universal legend in a plot and received the following result. I tried using "layout" but I can't seem to work on the "empty" plot (where I have to have the legend). I tried "oma" but I couldn't improve the quality of the plot, and that I didn't know how to specify all the line types using the keyboard, infact it is line type 4 (I had _._ and it didn't look nice). # I am trying to create a graph with 6 panels, but would like to have a universal legend as each panel merely denotes a separate stratum. The legend has to be at the bottom. I use "par(mfrow=c(2,3))" to get the panels, but am not sure how to put the legend below the whole graph. Thanking you as always J.R. Lockwood I don't think you can do this without layout() which you can use to create a separate graphical area within the plot region, where you can put the legend. i could be wrong though Peter Dalgaard You need to look at mtext(,outer=TRUE), plus par(oma=) to make room for the text in the outer margins. Try layout() for finer adjustments (and some drawbacks). See ?layout how to set up a 7th figure (legend) of full with below the desired six other figures. Uwe Ligges __ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Universal legend for graph...
Dear R-Users: I am trying to create a graph with 6 panels, but would like to have a universal legend as each panel merely denotes a separate stratum. The legend has to be at the bottom. I use "par(mfrow=c(2,3))" to get the panels, but am not sure how to put the legend below the whole graph. Thanking you as always. Vumani Dlamini Swaziland _ Add photos to your e-mail with MSN 8. Get 2 months FREE*. __ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] negatively correlated binary data
Dear R users: Is there a way to create negatively correlated binary data in R. I have looked at bindata, but it seems it only works with positively correlated data. Thanks. Vumani Dlamini Swaziland __ [EMAIL PROTECTED] mailing list http://www.stat.math.ethz.ch/mailman/listinfo/r-help
