[R] deal with R.package panel
hello! my question conserns with use of panel package (written by R.C.Gentlman) (unfortunately the manual and help sites are very short) 1. is it possible to do analysis just without a(ny) covariate? i suggest do it by introducing a covariate with level=0 in all obervations, this because of Q(z)=Q_o exp(beta*z), but it seemingly doesn't work 2. in the option gamma in the call of panel function: do you mean an initial value for parameter vector gamma? say if i have 3 theta-parameters, so i have to initialize gamma=c(xxx,xxx,xxx), correct? 3. are the (first ) observed times =0 allowed (in $time vectors) or schould in such a case begin with =1, if there are any? thanks for your response __ 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] cannot.allocate.memory.again and 32bit---64bit
hello! -- i use 32bit.Linux(SuSe)Server, so i'm limited with 3.5Gb of memory i demonstrate, that there is times to times a problem with allocating of objects of large size, for example 0.state (no objects yet created) gc() used (Mb)gc trigger (Mb)max used (Mb) Ncells 162070 4.4 35 9.4 35 9.4 Vcells 59921 0.5 786432 6.0 281974 2.2 1.state: let create now a vector of large size -- my.vector-rnorm(10*500) object.size(my.vector)/1024^2 [1] 381.4698 10*500*8/1024^2 #calculate object.size directly [1] 381.4697 gc() used (Mb) gc trigger (Mb) max used (Mb) Ncells 162257 4.4 35 9.4 35 9.4 Vcells 50060239 382.0 50412232 384.7 50060419 382.0 3.state: well, let create a matrix of the same size from this vector -- my.matrix-matrix(my.vector,nrow=10,ncol=500) gc() used (Mb) gc trigger(Mb) max used(Mb) Ncells 162264 4.4 359.4 359.4 Vcells 100060241 763.4150315042 1146.9 150060261 1144.9 object.size(my.matrix)/1024^2#calculate object.size directly [1] 381.4698 so, the matrix actually - according to the used.Mb - needs the same Mb as the vector. but, the trigger.Mb - and i still have problems with understanding of this - grows ennormously. and i can sure, i had received the cannot allocate the vector of xxxKb-error last time, trying the same experiment. if we know, that the matrix (or array generally) is acctually alloccated as a vector (with removed dimensions), why do we need so much trigger.Mb for it? is it a problem for R only on a 32bit? what is the difference with recpect to trigger.memory if i use 64bit (i didn't yet)? thanks for your advice -- __ 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] gc() and gc trigger
hello, the question concerning to the memory used and g.c. after having removed objects. What is wrong? bevor --- gc() used (Mb) gc trigger (Mb) max used (Mb) Ncells 313142 8.4 1801024 48.1 1835812 49.1 Vcells809238 6.2 142909728 1090.4 178426948 1361.3 hier all attached objects detached and other no more used ones removed; also the objects which could change their size (like .Traceback) are checked; nothing unusual - after -- for (i in 1:30) gc() gc() used(Mb) gc trigger (Mb) max used (Mb) Ncells 3131498.4 1152655 30.8 1835812 49.1 Vcells 8092616.2 3218039 19.7 178426948 1361.3 object.size(mget(ls(all=T),envir=.GlobalEnv)) / 1024^2 [1] 9.829926 N.B.!!! the max used is not put back q() after having restarted R-prozess - gc() used(Mb) gc trigger (Mb) max used (Mb) Ncells 3024978.1 46787512.5 35 9.4 Vcells 7853466.0 1193335 9.2 9236127.1 thnks for hint ps: some more understandable information about how R manages the memory and to the output of gc(), specially about gc trigger would be helpful __ 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 --max-vsize=4000M --max-nsize=4000M
hello! what's wrong??? i use 32 bit machine. every time i 'm connecting the server to do my work with R on it , after having done R --max-vsize=4000M --max-nsize=4000M i become wrong limits specification for Ncells: . gc() used(Mb) gc trigger (Mb) limit (Mb) Ncells 698799 18.7 1644099 44 112000 Vcells 65744346501.6 189257063 1444 4000 is there anything wrong? thanks a lot __ 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] lme4---GLMM
Hello, I'm very sorry for my repeated question, which i asked 2 weeks ago, namely: i'm interested in possibly simple random-part specification in the call of GLMM(...) (from lme4-package) i have a random blocked structure (i.e. ~var.a1+var.a2+var.a3, ~var.b1+var.b2,~var.c1+var.c2+var.c3+var.c4), and each one part of it i would like to model as Identity-structure matrix. So i had, in symbols of nlme-package, and for only one cluster-variable my.Subject: random=list(my.Subject=pdBlocked(list(pdIdent(~var.a1+var.a2+var.a3,...),pdIdent(~var.b1+var.b2,...),pdIdent(~var.c1+var.c2+var.c3+var.c4),...))) As the lme4-package doesn't use the pdMat-classes for specification of the random-part in GLMM,i used, with advice of Douglas Bates, the explicit specification in the GLMM-call (this call can also been simplified, if i first attach the lme4 package and then the nlme-package): GLMM(., random=list(my.Subject=~var.a1+var.a2+var.a3, my.Subject=~var.b1+var.b2, my.Subject=var.c1+var.c2+var.c3+var.c4),.) and really could make estimates in such a way. The problem is how can i specify a simple matrix sructure for each block, such as pdIdent(...) in symbols of lme4. is it possible? Does the lme4 any use of pdMat-specification or something like this? If not, it seems, that i actually specify a general Covariance Matrix for each block, what i in principle don't want. thanks in advance __ 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] lme4 -- GLMM
hello! this is a question, how can i specify the random part in the GLMM-call (of the lme4 library) for compound matrices just in the the same way as they defined in the lme-Call (of the nlme library). For example i would just need random=list(my.Subject=pdBlocked(list(pdIdent(~... , ...),pdIdent(~... , ... this specification , if i also attach library(nlme) , is not accepted in the GLMM-call, though the simple form random=list(my.Subject=pdIdent(~...,...)) is accepted. what is the analogous of pdBlocked Co from nlme in lme4? thanks for replay __ 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] 2: lme4 --- GLMM
Douglas Bates wrote: The GLMM function in the lme4 package allows you to specify crossed random effects within the random argument without the need for the pdBlocked and pdIdent constructions. Simply ensure that your grouping factors are defined in such a way that each distinct group has a different level in the grouping factor (this is usually not a problem for crossed grouping factors but can be a problem with nested factors) and list them. For example random = list(rows = ~ 1, columns = ~1) the reason is that i actually want to use just one group. factor with two and more blocks, each of them would have the simple diag. structure, just as was possible with like pdIdent(...,...,...) specification in nlme-package. i also wanted to give initial values via value=, so i would really need to define i.e.: random=list(my.Subject=pdBlocked(pdIdent(value=labda1, form=~var.a1+...+var.am, nam=...), pdIdent(value=labda2, form=~var.b1+...+var.bn, nam=...),...)) or just, because pdBlocked is not accepted random=list(my.Subject=list(pdIdent(value=labda1, form=~var.a1+...+var.am, nam=...), pdIdent(value=labda2, form=~var.b1+...+var.bn, nam=...),...)) or for last, if really were possible, alternatively without having attached nlme: random=list( form=~var.a1+...+var.am | my.Subject), form=~var.b1+...+var.bn | my.Subject , ... ) but each time i recieve Error in switch(mode(x), NULL = structure(NULL, class = formula), : invalid formula so i don't know, how it can go. the reason for trying to use the lme4 is its sparse-matrix orientation and so comutationally more efficient. thank you for hint in advance __ 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
(Re: [R] lme in R-2.0.0: Problem with lmeControl) and parameter specification
Hello! Thanks a lot to Douglas Bates for your advice. concerning the lme(...) function i wanted to put four other questions. 1.in the specification of initial values in the pdMat-constructor i probably define a standard deviation (sigma_b) and not a variance (sigma_b^2). For instance in the Rail example in Pinheiro/Bates on p.81 if i specify a random parameter as random=pdIdent(value=lambda-diag(1000,1),form=~1), (in S-plus), then the call to the lme(...) with just 0 iterations produces: fm1Rail.lme-lme(travel~1,data=Rail,random=pdIdent(value=lambda-diag(1000,1),form=~1),control=list(msMaxIter=0,msVerbose=TRUE,niterEM=0)) Iteration: 0 , 1 function calls, F= 66.37359 Parameters: [1] -3.453878 Warning messages: ITERATION LIMIT REACHED WITHOUT OTHER CONVERGENCE in: ms( ~ - logLik(lmeSt, lmePars), start = list(lmePars = c(coef(lmeSt))), if i now print out an estimated std.dev. for sigma_b i get: (fm1Rail.lme$sigma)*exp(unlist(fm1Rail.lme$modelStruct)) reStruct.Rail 107.6767 lambda [,1] [1,] 1000 so that the estimated variance would be 107.6767 ^2 = 11594.27 what is much grater then 1000. But hier we know that with iterations the value of variance will reduce ( and at the convergence the StdDev is 24.80547 ) so i think that lambda=1000 is specified equal to sigma_b as initial value. 2.What is the meaning for 0-Iteration? 3.are the parametersfixed=beta and random=sigmabeing calculated (just on time) only after all iterations have run, or they also be updated at every iteration with new value of teta ? if the latter, how can i get them for each run ? 4.Is it in principle possible to hold a variance components parameter, say sigma_b, as in Rail-example, fixed (on specified value) through all the iteration steps (without changing it) and only optimize for teta=log(sigma_b/sigma_epsilon) with fixed known value of sigma_b ? how can it be done ? Thank you for replay __ [EMAIL PROTECTED] 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] lme in R-2.0.0: Problem with lmeControl
Hello! One note/question hier about specification of control-parameters in the lme(...,control=list(...)) function call: i tried to specify tne number of iteration needed via lme(,control=list(maxIter=..., niterEM=...,msVerbose=TRUE)) but every time i change the defualt values maxIter (e.g. maxIter=1, niterEM=0) on ones specified by me, the call returns all the iterations needed until it's converged. and this is exactly the problem i will to get round. (e.g. in example on p.81 of Pinheiro/Bates,2000: fm1Rail.lme-lme(...,control=list(maxIter=1,...))) so i have tried with option msMaxIter=... and this works. The other problem is, that i even can not see (in R !!!) the output from iterations, despite the msVerbose=TRUE specification and setting options(verbose=TRUE) (The S-plus can do it but also ignoring the maxIter=... specification) Thank you for your hint __ [EMAIL PROTECTED] 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] seeking for the GLME-package (Jose Pinheiro)
Hello! could you give me some advice where i can finde out/recieve/download a GLME package, written by Jose Pinheiro. i couldn'ti find it in the package lists (probably becouse it is still in a beta version). thank you for your replay __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
