Re: [R] Mixed model
Thanks Most thoughtful... Regards S From: Spencer Graves [mailto:[EMAIL PROTECTED] Sent: Mon 27/06/2005 19:52 To: Stephen Cc: Douglas Bates; r-help@stat.math.ethz.ch Subject: Re: [R] Mixed model I often think carefully about what I want and store only that. For example, I might do something like the following: b1 - coef(lme(...)) kb - length(b1) B - array(NA, dim=c(nb, kb)) for(i in 1:nb){ B[i, ] - coef(lme(...)) } With fit[[i]] - lme(...), you store, as Doug said, a copy of the model frame (the parts of Dataset that are needed to evaluate the model) plus a lot of other information) for each pass through the loop. Since you are doing a simulation, you probably only really care about a few numbers for each i. Identify those and only store those each time through the loop. spencer graves p.s. Have you considered simulate.lme? Stephen wrote: Hi Thank you for your comments. Yes you are correct its a very big data set. Perhaps I am best splitting it up and then importing to R. The reason for the loop is that I am conducting the equivalent of Split file in SPSS. Specifically, I am conducting the analysis for each value of on the grouping variable 'runnb'. If there is a less memory intensive way of doing this I'd appreciate knowing about it. Many Thanks and comments appreciated Regards Stephen From: Douglas Bates [mailto:[EMAIL PROTECTED] Sent: Sun 26/06/2005 17:01 To: Stephen Cc: r-help@stat.math.ethz.ch Subject: Re: [R] Mixed model On 6/26/05, Stephen [EMAIL PROTECTED] wrote: Hi All, I am currently conducting a mixed model. I have 7 repeated measures on a simulated clinical trial. If I understand the model correctly, the outcome is the measure (as a factor) the predictors are clinical group and trial (1-7). The fixed factors are the measure and group. The random factors are the intercept and id and group. I tried using 2 functions to calculate mixed effects. Following previous correspondence . Dataset - read.table(C:/Program Files/R/rw2011/data/miss/model1a.dat, header=TRUE, sep=\t, na.strings=NA, dec=., strip.white=TRUE) attach(Dataset) require (nlme) with(Dataset, table(runnb, id, grp)) b.lvls - table(Dataset$runnb) nb - length(b.lvls) fit - vector(mode=list, nb) for(i in 1:nb) fit[[i]]- lme (trans1 ~ Index1 + grp, random = ~ 1 | id / grp , data = Dataset, na.action = na.exclude) This (above) worked OK only I am having memory problems. I have a gig of RAM set at --sdi --max-mem-size=512M (complete version below) I am wondering if running the file as a database be slower / faster? Then I read that lme4 does it quicker and more accurately so I thought that I should re-run the code but from the for line: for (i in 1:nb) + fit[[i]] - lmer(trans1 ~ Index1 + grp + (1|id:grp) + (1|id), + Dataset, na.action = na.exclude) Producing Error in lmer(trans1 ~ Index1 + grp + (1 | id:grp) + (1 | id), Dataset, : flist[[2]] must be a factor of length 20 In addition: Warning messages: 1: numerical expression has 20 elements: only the first used in: id:grp 2: numerical expression has 20 elements: only the first used in: id:grp Check str(Dataset) and, if necessary, convert id to a factor with Dataset$id - factor(Dataset$id) In is not surprising that you are running into memory problems. Look at the size of one of the fitted objects from lme or from lmer. They are very large because they contain a copy of the model frame (the parts of Dataset that are needed to evaluate the model) plus a lot of other information. You have a large Dataset and you are saving multiple copies of it although I must admit that I don't understand why the calls to lme or lmer are in a loop. ?? http://mail.nana.co.il [[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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 ?? http://mail.nana.co.il [[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
Re: [R] How to convert c:\a\b to c:/a/b?
This is based on a false premise. R _can_ read files containing such names: use readLines() or scan(allowEscapes=FALSE). Someone else wrote the C for you! The restriction is that if you pass character strings to the parser, they are interpreted according to the documented rules, including interpreting escapes. On Tue, 28 Jun 2005, Mulholland, Tom wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Henrik Bengtsson Sent: Tuesday, 28 June 2005 2:54 AM To: Spencer Graves Cc: r-help@stat.math.ethz.ch; Dirk Eddelbuettel Subject: Re: [R] How to convert c:\a\b to c:/a/b? ... snipped Thus, you cannot write your program such that it fools the parser, because your program is evaluated first after the parser. In other words, there is no way you can get nchar(\n) to equal 2. I had been waiting for this answer because it was the conclusion I had come to. Given that I mainly work in a windows world this has been a problem. For various reasons I receive files liberally sprinkled with such pathnames. I generally pre-process them using whatever is at hand. It's not a big problem, just annoying to have to explain to collegues that this is something R can't do. Not a good advertisment for those who have no idea about escape codes. However I can't believe that this problem cannot be solved. The thoughts that have come through my head are to write a c routine that effectively ignores the possibility that \n means newline and thus remaps all the escape codes into text (\\ and the character code.) I've never written in C which is one of the reasons that I have never attempted this. I would be interested in any thoughts about the viability of my proposal. It seems an awful lot of work (at least for someone who hasn't done this sort of stuff before) for something that can be achieved in many other ways. -- 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
Re: [R] GetRNGstate() crashes in Windows
On Tue, 28 Jun 2005, Ju-Sung Lee wrote: I think I've misread the intent of the GetRNGstate() and other API functions; these only work when embedded in C functions eventually called from R. Apologies for the mistake. First, the posting guide asks you to send programming questions to the R-devel list. You can call some of the API functions from a C program, but you need to build a standalone libRmath: see src/nmath/standalone/README in the R sources. But GetRNGstate is not one of them, and `Writing R Extensions' states clearly which they are. See its section `Standalone Mathlib1'. Linking like gcc -o prog.exe prog.c -I{R's include path} R.dll is not supported by MinGW (even though it is documented, it fails far too often and the reply to my bug report was that it was not supposed to work!). On Tue, 28 Jun 2005, Ju-Sung Lee wrote: Hi, Has anyone managed to successfully call GetRNGstate() / PutRNGstate() without crashing in a Windows environment (spec. XP)? I've compiled successfully using both the latest Cygwin, latest Mingw, and the version of Mingw suggested in Building R for Windows website, but when the executable runs, it crashes; the functions themselves can run when I omit GenRNGstate()/PutRNGstate, but return the same values every time. My compilation command is, having copied R.dll to my current directory: gcc -o prog.exe prog.c -I{R's include path} R.dll I've also tried: gcc -o prog.exe prog.c -I{R's include path} -L./ -lR I've managed to successfully compile Rmath.dll and call R functions that way, but remain puzzled why the first way doesn't work. I've read mention of initializing R (within a C program I assume) but haven't found the documentation that explains this. My test program is simply: #include R.h int main() { GetRNGstate(); PutRNGstate(); return 0; } I've also tried #include-ing host of the other .h files. Thanks! -- 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] Help
Hi, Please, can you tell if there are any package or R code for STAR models estimation and test misspecification. Thanks in advance Best regards AJMI Noomen PHD Student TUNISIA __ 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] How to convert c:\a\b to c:/a/b?
Mulholland, Tom wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Henrik Bengtsson Sent: Tuesday, 28 June 2005 2:54 AM To: Spencer Graves Cc: r-help@stat.math.ethz.ch; Dirk Eddelbuettel Subject: Re: [R] How to convert c:\a\b to c:/a/b? ... snipped Thus, you cannot write your program such that it fools the parser, because your program is evaluated first after the parser. In other words, there is no way you can get nchar(\n) to equal 2. I had been waiting for this answer because it was the conclusion I had come to. Given that I mainly work in a windows world this has been a problem. For various reasons I receive files liberally sprinkled with such pathnames. I generally pre-process them using whatever is at hand. It's not a big problem, just annoying to have to explain to collegues that this is something R can't do. Not a good advertisment for those who have no idea about escape codes. Please, note this basically only applies to source(), expressions at the R prompt (and unfortunately read.table()), and therefore you should not have to pre-process you files. Here are some illustrating example. It is a good exercise to convince yourself that you understand why you get the different results; code - x - \D:\\spencerg\\statmtds\\R\\Rnews\ cat(file=foo.R, code) file.show(foo.R) # x - D:\spencerg\statmtds\R\Rnews x - NA eval(parse(text=code)) print(x) rm(x) [1] D:spencergstatmtdsRRnews source(foo.R) print(x) [1] D:spencergstatmtdsRRnews print(readLines(foo.R)) [1] x - \D:\\spencerg\\statmtds\\R\\Rnews\ print(scan(foo.R, what=character(0), allowEscapes=FALSE)) [1] x- [3] D:spencergstatmtdsRRnews print(read.table(foo.R)) V1 V2 V3 1 x - D:spencergstatmtdsRRnews print(readChar(foo.R, nchar=256)) [1] x - \D:\\spencerg\\statmtds\\R\\Rnews\ print(readBin(foo.R, what=integer(0), size=1, n=256)) [1] 120 32 60 45 32 34 68 58 92 115 112 101 110 99 101 114 103 92 115 [20] 116 97 116 109 116 100 115 92 82 92 82 110 101 119 115 34 Comment/suggestion: It would be nice if read.table() would pass argument 'allowEscapes' (or just '...') to scan(). /Henrik However I can't believe that this problem cannot be solved. The thoughts that have come through my head are to write a c routine that effectively ignores the possibility that \n means newline and thus remaps all the escape codes into text (\\ and the character code.) I've never written in C which is one of the reasons that I have never attempted this. I would be interested in any thoughts about the viability of my proposal. It seems an awful lot of work (at least for someone who hasn't done this sort of stuff before) for something that can be achieved in many other ways. Tom __ 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-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] .ts
Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Jver [EMAIL PROTECTED] __ 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] help--package:MIX
The general location model is assumed: continuous variables are assumed normal conditional on the categorical variables. ft. -- Fernando TUSELLe-mail: Departamento de Econometría y Estadística [EMAIL PROTECTED] Facultad de CC.EE. y Empresariales Tel: (+34)94.601.3733 Avenida Lendakari Aguirre, 83 Fax: (+34)94.601.3754 E-48015 BILBAO (Spain)Secr: (+34)94.601.3740 __ 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] STAR models estimation with R
Hi, Can you tell me if there are an R package or code for STAR model estimation and test misspecification. If no, how i could do this. Thanks in advance Best regards AJMI Noomen Phd student TUNISIA - [[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] nonparametric 2way repeated-measures anova
Dear useRs is there any nonparametric test for the analysis of variance in a design with two within-factors (repeated measures on both factors)? Friedman is not appropriate here, therefore I am grateful for any alternative test. thanks for any hint cheers christoph -- __ 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] How to import data as numeric array?
Did some search but couldn't find useful result. I am trying to read a n*m dimension data with read.table, what i need is a numeric array, is there any efficient way to allow me get this array directly instead of a list? I tried to use as.array() to change the mode, but seems it doesn't work and i got this error message: Error in dimnames-.data.frame(`*tmp*`, value = list(function (M) : invalid dimnames given for data frame what's wrong with the dimension names , should i delete them? how can i do that? Thanks a lot for any help. tong wang __ 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-pkgs] New package `party': A Laboratory for Recursive Part(y)itioning
Dear useRs, a new package for tree-structured regression is available on CRAN. This package implements a unified framework for recursive partitioning which embeds tree-structured regression models into a well defined theory of conditional inference procedures. Stopping criteria based on multiple test procedures are implemented. The methodology is applicable to all kinds of regression problems, including nominal, ordinal, numeric, censored as well as multivariate response variables and arbitrary measurement scales of the covariates. Extensible functionality for visualizing tree-structured regression models is available. Best, Torsten __ Package: party Title: A Laboratory for Recursive Part(y)itioning Date: $Date: 2005/06/27 06:38:16 $ Version: 0.2-2 Author: Torsten Hothorn, Kurt Hornik and Achim Zeileis Maintainer: Torsten Hothorn [EMAIL PROTECTED] Description: Unbiased recursive partitioning in a conditional inference framework. Depends: R (= 2.0.1), survival, grid, modeltools, coin Suggests: ipred SaveImage: yes License: GPL Packaged: Sun Jun 12 10:38:21 2005; hothorn ___ R-packages mailing list [EMAIL PROTECTED] https://stat.ethz.ch/mailman/listinfo/r-packages __ 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] How to import data as numeric array?
Please consult the `R Data Import/Export Manual'. You could be using scan(), as in matrix(scan(some_file, 0), nrow=n, byrow=TRUE) On Tue, 28 Jun 2005, tong wang wrote: Did some search but couldn't find useful result. I am trying to read a n*m dimension data with read.table, what i need is a numeric array, is there any efficient way to allow me get this array directly instead of a list? I tried to use as.array() to change the mode, but seems it doesn't work and i got this error message: Error in dimnames-.data.frame(`*tmp*`, value = list(function (M) : invalid dimnames given for data frame what's wrong with the dimension names , should i delete them? how can i do that? I think you wanted as.matrix here, not as.array: probably the call to the latter was incorrect, but we were not shown it. -- 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
Re: [R] Mixed model
Hi, Your last email was excellent and I got it working using the following. It did the job extremely quickly, more so a commercial piece of software my university uses! #This works fine - Dataset - read.table(C:/Program Files/R/rw2011/data/miss/model1.dat, header=TRUE, sep=\t, na.strings=NA, dec=.,strip.white=TRUE) Dataset$id - factor(Dataset$id) attach(Dataset) names (Dataset) require (nlme) fit - vector(mode=list, 100) i = 0 for (i in 1:100) { Dataset2 - subset(Dataset, subset= i == runnb) summary (fit[[i]]- lme (trans1 ~ Index1 + grp, random = ~ 1 | id / grp , data = Dataset2, na.action = na.exclude) ) i =i +1 } Strangely enough I was less lucky with LMER. for (i in 1:100) + { Dataset2 - subset(Dataset, subset= i == runnb) + fit[[i]] - lmer(trans1 ~ Index1 + grp + (1|id:grp) + (1|id), Dataset2, na.action = na.exclude) + i =i +1 + } Error in lmer(trans1 ~ Index1 + grp + (1 | id:grp) + (1 | id), Dataset2, : flist[[2]] must be a factor of length 4000 In addition: Warning messages: 1: numerical expression has 4000 elements: only the first used in: id:grp 2: numerical expression has 4000 elements: only the first used in: id:grp Many thanks to you all for your assistance on this issue. Regards Stephen From: Douglas Bates [mailto:[EMAIL PROTECTED] Sent: Sun 26/06/2005 17:01 To: Stephen Cc: r-help@stat.math.ethz.ch Subject: Re: [R] Mixed model On 6/26/05, Stephen [EMAIL PROTECTED] wrote: Hi All, I am currently conducting a mixed model. I have 7 repeated measures on a simulated clinical trial. If I understand the model correctly, the outcome is the measure (as a factor) the predictors are clinical group and trial (1-7). The fixed factors are the measure and group. The random factors are the intercept and id and group. I tried using 2 functions to calculate mixed effects. Following previous correspondence . Dataset - read.table(C:/Program Files/R/rw2011/data/miss/model1a.dat, header=TRUE, sep=\t, na.strings=NA, dec=., strip.white=TRUE) attach(Dataset) require (nlme) with(Dataset, table(runnb, id, grp)) b.lvls - table(Dataset$runnb) nb - length(b.lvls) fit - vector(mode=list, nb) for(i in 1:nb) fit[[i]]- lme (trans1 ~ Index1 + grp, random = ~ 1 | id / grp , data = Dataset, na.action = na.exclude) This (above) worked OK only I am having memory problems. I have a gig of RAM set at --sdi --max-mem-size=512M (complete version below) I am wondering if running the file as a database be slower / faster? Then I read that lme4 does it quicker and more accurately so I thought that I should re-run the code but from the for line: for (i in 1:nb) + fit[[i]] - lmer(trans1 ~ Index1 + grp + (1|id:grp) + (1|id), + Dataset, na.action = na.exclude) Producing Error in lmer(trans1 ~ Index1 + grp + (1 | id:grp) + (1 | id), Dataset, : flist[[2]] must be a factor of length 20 In addition: Warning messages: 1: numerical expression has 20 elements: only the first used in: id:grp 2: numerical expression has 20 elements: only the first used in: id:grp Check str(Dataset) and, if necessary, convert id to a factor with Dataset$id - factor(Dataset$id) In is not surprising that you are running into memory problems. Look at the size of one of the fitted objects from lme or from lmer. They are very large because they contain a copy of the model frame (the parts of Dataset that are needed to evaluate the model) plus a lot of other information. You have a large Dataset and you are saving multiple copies of it although I must admit that I don't understand why the calls to lme or lmer are in a loop. ?? http://mail.nana.co.il [[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
Re: [R] .ts
On Tue, 28 Jun 2005 09:20:54 +0200 pir2.jv wrote: Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Look at the zoo package for time series with arbitrary time scale (e.g., Date, chron, POSIXct, ...). The zoo vignette also explains the relation to other irregular time series classes available in package its, fCalendar and tseries. Z Jver [EMAIL PROTECTED] __ 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-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] enhanced multidimensional scaling?
Dear R list Would anyone be able to tell me whether it is possible to do enhanced multidimensional scaling (enhanced MDS) in R? In other words, something that goes beyond cmdscale by iteratively improving the fit between observed dissimilarities and inter-object distances, using the KYST algorithm (Kruskal, 1964). I have found several implementations of non-metric MDS in various packages but nothing like what I have described above. Thanks in advance -- Karen Kotschy Centre for Water in the Environment University of the Witwatersrand Johannesburg P/Bag X3, Wits, 2050 Tel: +2711 717-6425 __ 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] .ts
On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 09:20:54 +0200 pir2.jv wrote: Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Look at the zoo package for time series with arbitrary time scale (e.g., Date, chron, POSIXct, ...). The zoo vignette also explains the relation to other irregular time series classes available in package its, fCalendar and tseries. However, this might be a regular time series. For an example, see example(beav1, package=MASS). This a series measured every 10 mins. -- 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
Re: [R] How to import data as numeric array?
maybe you can use the array function,like array(CO2,dim(CO2),list(rownames(CO2),colnames(CO2))) and matrix is just a specif type of array,so maybe you can use as.matrix as.matrix(CO2) the above tow,the CO2 is a data.frame which can use read.table to read in. the third way is:use the scan the read the data and use array to change the data to array. On Tue, 28 Jun 2005 01:05:23 -0700 tong wang [EMAIL PROTECTED] wrote: Did some search but couldn't find useful result. I am trying to read a n*m dimension data with read.table, what i need is a numeric array, is there any efficient way to allow me get this array directly instead of a list? I tried to use as.array() to change the mode, but seems it doesn't work and i got this error message: Error in dimnames-.data.frame(`*tmp*`, value = list(function (M) : invalid dimnames given for data frame what's wrong with the dimension names , should i delete them? how can i do that? Thanks a lot for any help. tong wang __ 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 -- Department of Sociology Fudan University,Shanghai Blog:http://sociology.yculblog.com __ 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] .ts
On Tue, 28 Jun 2005 10:53:11 +0100 (BST) Prof Brian Ripley wrote: On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 09:20:54 +0200 pir2.jv wrote: Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Look at the zoo package for time series with arbitrary time scale (e.g.,Date, chron, POSIXct, ...). The zoo vignette also explains the relation to other irregular time series classes available in package its, fCalendar and tseries. However, this might be a regular time series. For an example, see example(beav1, package=MASS). This a series measured every 10 mins. ...which you could handle by creating a regular zoo series (of class zooreg). Z -- 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] enhanced MDS
Hi again Sorry, in looking again at sammon and isoMDS I see that they seem to do exactly what I want, except that they are non-metric, which means, as I understand it, that they relate the rank orders of the variables rather than the actual distances. Could I use these non-metric MDS packages even if my distances are metric? Thanks Karen -- Karen Kotschy Centre for Water in the Environment University of the Witwatersrand Johannesburg P/Bag X3, Wits, 2050 Tel: +2711 717-6425 __ 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] .ts
On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 10:53:11 +0100 (BST) Prof Brian Ripley wrote: On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 09:20:54 +0200 pir2.jv wrote: Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Look at the zoo package for time series with arbitrary time scale (e.g.,Date, chron, POSIXct, ...). The zoo vignette also explains the relation to other irregular time series classes available in package its, fCalendar and tseries. However, this might be a regular time series. For an example, see example(beav1, package=MASS). This a series measured every 10 mins. ...which you could handle by creating a regular zoo series (of class zooreg). Yes, but the standard ts analysis functions are for objects of class ts, and those suffice for regular time series. Some work for class zooreg and some do not. (More would if zooreg had a tsp() method.) Z -- 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 -- 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
Re: [R] enhanced MDS
On Tue, 28 Jun 2005, Karen Kotschy wrote: Hi again Sorry, in looking again at sammon and isoMDS I see that they seem to do exactly what I want, except that they are non-metric, which means, as I understand it, that they relate the rank orders of the variables rather than the actual distances. Could I use these non-metric MDS packages even if my distances are metric? Yes. BTW, Sammon is not ordinal. What `non-metric' means depends on who is using the term. -- 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
Re: [R] solving equation system
HI ALL i would like to solve a complex set of equations. i have four parameters and four equations. i could set up more equations since they are derived from the momnets of a particular distribution. the parameters are NON LINEAR!!! AND the eqautions are of the form: phi(i)=function(a,x,y,z) is there a package or group of commands that might be used in order to solve the system directly? thanking you in advance / allan Spencer Graves wrote: Have you considered writing a function to compute the sum of squares of deviations from equality and using optim? I use sum of squares not sum of absolute values, because if my functions are differentiable, the sum of squares will also be differentible while the sum of absolute values will not be. This means that sum of absolute values will not work well with a quasi-Newton algorithm. Also, have you considered making plots? If I understand your example, you can solve for lambda using (II) as lambda = x/mean(X). Then you can use (I) to solve for c. To understand this, it would help to plot the digamma function. If you do this (e.g., http://mathworld.wolfram.com/DigammaFunction.html), you will see that there are countably infinite solutions to this equation. If you want the positive solution, I suggest you try to solve for ln.c = log(c) rather than c directly, because that should make optim more stable. More generally, it often helps to make, e.g., contour or perspective plots and to try to find a parameterization that will make the sum of squares of errors approximatly parabolic in your parameters. My favorite reference on this is Bates and Watts (1988) Nonlinear Regression Analysis and Its Applications (Wiley). There may be better, more recent treatments of this subject, but I am not familiar with them. spencer graves p.s. I never (no never, not ever) use c as a variable name, because it is the name of a common R function. R is smart enough to distinguish between a function and a non-function in some contexts but not in all. When I want a name for a new object, I routinely ask R to print my proposed name. If it returns Error: object ... not found, I can use Carsten Steinhoff wrote: Hello, I want to solve some two dimensional equation system with R. Some systems are not solvable analytically. Here is an example: (I)1/n*sum{from_i=1_to_n}(Xi) = ln lambda + digamma(c) (II)mean(X) = x / lambda I want to find lambda and c, which R-function could do that task? Carsten [[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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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-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-postgresql
hello where can download R-postgresql from ? regards Ahmet Temiz __ XamimeLT - installed on mailserver for domain @deprem.gov.tr Queries to: [EMAIL PROTECTED] __ The views and opinions expressed in this e-mail message are ...{{dropped}} __ 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] R-postgresql
orkun wrote: hello where can download R-postgresql from ? Please read the posting guide. What is R-postgresql? For which version of R on which platform? The following two web pages might help you to precise your question: http://www.omegahat.org/RSPostgres/ http://rpgsql.sourceforge.net/ Uwe Ligges regards Ahmet Temiz __ XamimeLT - installed on mailserver for domain @deprem.gov.tr Queries to: [EMAIL PROTECTED] __ The views and opinions expressed in this e-mail message are ...{{dropped}} __ 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-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] add transformed columns into a matrix
Supposing 'inmatrix' is a matrix with coloumn names 'x1', 'x2' and 'x3'; how about something like model.matrix(~ (x1 + x2 + x3)^2 + log(x1) + log(x2) + log(x3) + sqrt(x1) + sqrt(x2) + sqrt(x3) - 1, as.data.frame(inmatrix)) -- Bjørn-Helge Mevik __ 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] nonparametric 2way repeated-measures anova
Cristoph: Take a look on Brunner, E., Domhof, S. and Langer, F. (2002). Nonparametric analysis of longitudinal data in factorial experiments. John Wiley Sons, New York. for the theory. The authors have R and SAS code available on http://www.ams.med.uni-goettingen.de/de/sof/ld/makros.html Sincerely, -- Frederico Zanqueta Poleto [EMAIL PROTECTED] -- An approximate answer to the right problem is worth a good deal more than an exact answer to an approximate problem. J. W. Tukey Christoph Lehmann wrote: Dear useRs is there any nonparametric test for the analysis of variance in a design with two within-factors (repeated measures on both factors)? Friedman is not appropriate here, therefore I am grateful for any alternative test. thanks for any hint cheers christoph -- __ 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 -- Frederico Zanqueta Poleto [EMAIL PROTECTED] -- An approximate answer to the right problem is worth a good deal more than an exact answer to an approximate problem. J. W. Tukey __ 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] index of dispersion
Hi list I'm looking for statisitic measurements describing the pattern of points within a given polygon. Is there a function calculating the Index of Dispersion and/or are there other functions summarising an observed pattern? Markus Schwarz . Markus Schwarz Wissenschaftliche Mitarbeiterin Eidg. Forschungsanstalt WSL Forschungsprogramm Musterland Zürcherstrasse 111 CH-8903 Birmensdorf Telefon +41-44-739 22 87 Fax +41-44-739 22 15 [EMAIL PROTECTED] http://www.wsl.ch/staff/markus.schwarz/ . __ 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] Programming Books
I am about to embark on the writing of a program for R and I wanted to know a good reference book that would be useful. There has been some talk about S programming the Springer publication ISBN:0387989668. Is that what most people are using or are there other books that might be dedicated to R. Philip Bermingham __ 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] How to convert c:\a\b to c:/a/b?
On 27-Jun-05 Spencer Graves wrote: Thanks, Dirk, Gabor, Eric: You all provided appropriate solutions for the stated problem. Sadly, I oversimplified the problem I was trying to solve: I copy a character string giving a DOS path from MS Windows Explorer into an R script file, and I get something like the following: D:\spencerg\statmtds\R\Rnews I want to be able to use this in R with its non-R meaning, e.g., in readLine, count.fields, read.table, etc., after appending a file name. Your three solutions all work for my oversimplified toy example but are inadequate for the problem I really want to solve. The various responses show that solving this problem directly within R may be, well, problematic! I'm not a perl user, but possibly Spencer Graves's speculation might eventually be brought into a straightforward solution. Unfortunately, for this query, I'm not a Windows users either, so can't be authoritative about how practical the following may be for Spencer's problem (and similar). In the Unix world, the string editor program 'sed' simply mops up problems of this kind. For example, I just did: echo D:\spencerg\statmtds\R\Rnews | sed '[EMAIL PROTECTED]@/@g' and got the response: D:/spencerg/statmtds/R/Rnews Note: the parsing of the 'sed' command is as follows: [EMAIL PROTECTED]@[EMAIL PROTECTED] means substitute y for every (g = global) occurrence of x. The character to be replaced (x=\) is the escape character so needs to be escaped (\\); but apart from this it's straightforward. The usual separator is / instead of @, but I used @ to simplify things since the substitute itself is y=/. An alternative using / as separator would be echo D:\spencerg\statmtds\R\Rnews | sed 's/\\/\//g' D:/spencerg/statmtds/R/Rnews which is a bit more complicated but still straightforward (depending on your eyesight). Here, both x=\ and y=/ need to be escaped. 'sed', and a lot of other useful stuff, is available as GNU tools which can be installed on Windows, and allow this kind of thing to be very slickly done, but outside of R of course. I also make much use of 'awk' (but you can equally use perl) for tidying up CSV files exported from Excel, and for all sorts of rearrangements and substitutions in data files. While Unix users take this sort of thing for granted, I suggest to Windows users that it could be well worth installing the GNU tools, since they are very useful indeed. It's not clear from Spencer's follow-up whether doing this kind of preliminary work outside of R, and then bringing the results into R (e.g. via the clipboard or a file) is practical in his context. If not (i.e. all the work has to be done inside R), then of course my sugestion above is not helpful in this case! Best wishes, Ted. E-Mail: (Ted Harding) [EMAIL PROTECTED] Fax-to-email: +44 (0)870 094 0861 Date: 28-Jun-05 Time: 13:38:37 -- XFMail -- __ 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] Programming Books
Philip Bermingham wrote: I am about to embark on the writing of a program for R and I wanted to know a good reference book that would be useful. There has been some talk about S programming the Springer publication ISBN:0387989668. Is that what most people are using or are there other books that might be dedicated to R. Yes, this is a ggod book, quite a lot of other books, online material and paper are mentioned on CRAN, e.g. CRAN/other-docs.html the latter points to, e.g., http://www.r-project.org/other-docs.html and http://www.r-project.org/doc/bib/R-publications.html Uwe Ligges Philip Bermingham __ 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-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] function for cumulative occurrence of elements
Hello, I have a data set with 9700 records, and 7 parameters. The data were collected for a survey of forest communities. Sample plots (1009) and species (139) are included in this data set. I need to determine how species are accumulated as new plots are considered. Basically, I want to develop a species area curve. I've included the first 20 records from the data set. Point represents the plot id. The other parameters are parts of the information statistic H'. Using Table, I can construct a data set that lists the occurrence of a species at any Point (it produces a binary 0/1 data table). From there it get confusing, regarding the most efficient approach to determining the addition of new and or repeated species occurrences. ptcount - table(sppoint.freq$species, sppoint.freq$Point) From here I've played around with colSums to calculate the number of species at each Point. The difficulty is determining if a species is new or repeated. Also since there are 1009 points a function is needed to screen every Point. Two goals are of interest: 1) the species accumulation curve, and 2) an accumulation curve when random Points are considered. Any help would be greatly appreciated. Thank you Steve Friedman Pointspecies frequency point.list point.prop log.prop point.hprime 1 7 American elm 7 27 0.25925926 -1.3499267 0.3499810 2 7 apple 2 27 0.07407407 -2.6026897 0.1927918 3 7 black cherry 8 27 0.29629630 -1.2163953 0.3604134 4 7 black oak 1 27 0.03703704 -3.2958369 0.1220680 5 7chokecherry 1 27 0.03703704 -3.2958369 0.1220680 6 7 oak sp 1 27 0.03703704 -3.2958369 0.1220680 7 7 pignut hickory 1 27 0.03703704 -3.2958369 0.1220680 8 7 red maple 1 27 0.03703704 -3.2958369 0.1220680 9 7 white oak 5 27 0.18518519 -1.6863990 0.3122961 10 9 black spruce 2 27 0.07407407 -2.6026897 0.1927918 11 9blue spruce 2 27 0.07407407 -2.6026897 0.1927918 12 9missing12 27 0. -0.8109302 0.3604134 13 9 Norway spruce 8 27 0.29629630 -1.2163953 0.3604134 14 9 white spruce 3 27 0. -2.1972246 0.2441361 1512 apple 2 27 0.07407407 -2.6026897 0.1927918 1612 black cherry 1 27 0.03703704 -3.2958369 0.1220680 1712 black locust 1 27 0.03703704 -3.2958369 0.1220680 1812 black walnut 1 27 0.03703704 -3.2958369 0.1220680 1912 lilac 3 27 0. -2.1972246 0.2441361 2012missing 2 27 0.07407407 -2.6026897 0.1927918 __ 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] STAR models estimation with R
The only STAR model I have come across is Smooth Threshold Autoregressive time series model, see Tong, Non Linear Time Series. I am not aware of any package that has implemented threshold models. Because statistical techniques are used widely across many different disciplines it is inevitable that naming conventions will diverge, so the same technique may have different names in different areas of study. Perhaps it should be added to the posting guide that most general name of a technique should be used, and in the case of more obscure techniques a brief reference to a standard text or paper. Hamilton, see Time Series Analysis, uses the EM algorithm to estimate such models so it should be possible to do in R. HTH Phineas Campbell -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of noomen ajmi Sent: Tuesday, June 28, 2005 8:44 AM To: r-help@stat.math.ethz.ch Subject: [R] STAR models estimation with R Hi, Can you tell me if there are an R package or code for STAR model estimation and test misspecification. If no, how i could do this. Thanks in advance Best regards AJMI Noomen Phd student TUNISIA - [[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-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] where can i download the metrics package?
i learn it from metrics: Towards a package for doing econometrics in Rbut i can not find it in cran. -- Department of Sociology Fudan University,Shanghai Blog:http://sociology.yculblog.com __ 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] where can i download the metrics package
Look for package Ecdat. __ 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 data frames for EDA: Insert, Change name, delete columns? (Newcomer's question)
I am finding complex analyses easier than some elementary operations in R. In particular I want to do some low level exploratory data analyses with data in a data frame but cannot find commands to easily insert, remove (delete), rename, and re-order (arbitrarily, not sort) columns. I see that the micEcon package has an insertCol command, but that is for matrices, not data frames. I have looked through several introductory texts but all seem to stop short of explaining the commands needed for the above. Could a reader provide a reference where such commands are documented? Thank you, Ben Fairbank [[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
Re: [R] STAR models estimation with R
It is likely that here STAR means the Smooth Transition Autoregressive model which generalizes the Threshold Autoregressive (TAR) model to allow for gradual switching between regimes. The model can be estimated in R by applying maximum likelihood (ML) estimation or Monte Carlo Markov Chain (MCMC) estimation. Use Google to find related papers. Regards, Hannu Kahra -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Phineas Campbell Sent: 28. kesäkuuta 2005 15:11 To: r-help@stat.math.ethz.ch Subject: Re: [R] STAR models estimation with R The only STAR model I have come across is Smooth Threshold Autoregressive time series model, see Tong, Non Linear Time Series. I am not aware of any package that has implemented threshold models. Because statistical techniques are used widely across many different disciplines it is inevitable that naming conventions will diverge, so the same technique may have different names in different areas of study. Perhaps it should be added to the posting guide that most general name of a technique should be used, and in the case of more obscure techniques a brief reference to a standard text or paper. Hamilton, see Time Series Analysis, uses the EM algorithm to estimate such models so it should be possible to do in R. HTH Phineas Campbell -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of noomen ajmi Sent: Tuesday, June 28, 2005 8:44 AM To: r-help@stat.math.ethz.ch Subject: [R] STAR models estimation with R Hi, Can you tell me if there are an R package or code for STAR model estimation and test misspecification. If no, how i could do this. Thanks in advance Best regards AJMI Noomen Phd student TUNISIA - [[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-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-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] factor to character
Using RODBC, when I select from a table strings (chars and varchars) come as factors. What is the best way, speed wise, to convert these columns back to strings (perhaps using as.character). __ 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] .ts
R News (www.r-project.org - Documentation: Newsletter, (4/1) June 2004, pp. 29-32, and (2/2) June 2002, pp. 2-10 contain excellent articles on date-time classes and time series capabilities in R. In addition, I have found ch. 14 in Venables and Ripley (2002) Modern Applied Statistics in S, 4th ed. (Springer) and the vignette for the zoo package to be excellent. Greater flexibility in modeling and analysis are provided with the Rmetrics project (www.rmetrics.org or http://www.itp.phys.ethz.ch/econophysics/R/), but that also requires greater effort to learn. spencer graves Prof Brian Ripley wrote: On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 10:53:11 +0100 (BST) Prof Brian Ripley wrote: On Tue, 28 Jun 2005, Achim Zeileis wrote: On Tue, 28 Jun 2005 09:20:54 +0200 pir2.jv wrote: Goog morning! How to construct a time serie in hours or in minuts, ... The examples are all in years Excuse for questions so basic! Look at the zoo package for time series with arbitrary time scale (e.g.,Date, chron, POSIXct, ...). The zoo vignette also explains the relation to other irregular time series classes available in package its, fCalendar and tseries. However, this might be a regular time series. For an example, see example(beav1, package=MASS). This a series measured every 10 mins. ...which you could handle by creating a regular zoo series (of class zooreg). Yes, but the standard ts analysis functions are for objects of class ts, and those suffice for regular time series. Some work for class zooreg and some do not. (More would if zooreg had a tsp() method.) Z -- 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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] R demos
Federico Calboli wrote: Hi All, I am currently preparing some form of slideshow introducing R and its capabilities for some colleagues. The thing will be about 30 mins, and I'd like to have some pretty pictures and some amazing facts (I'm trying to sell, obviously :)). Can I ask if it's possible to easily retrieve a gross figure of the number of functions in R considering the base install and all the libraries available? Apart from graphics and lattice, are there any more packages producing eye catching graphics (possibly with a survival analysis/epidemiological bend)? Cheers, Federico Calboli In the package relax you find the function slider(). The help page of slider shows a nice application: R.veil.in.the.wind() Here are the definitions of slider and R.veil.in.the.wind: # definition of slider slider-function (sl.functions, sl.names, sl.mins, sl.maxs, sl.deltas, sl.defaults, but.functions, but.names, no, set.no.value, obj.name, obj.value, reset.function, title) { if (!missing(no)) return(as.numeric(tclvalue(get(paste(slider, no, sep = ), env = slider.env if (!missing(set.no.value)) { try(eval(parse(text = paste(tclvalue(slider, set.no.value[1], )-, set.no.value[2], sep = )), env = slider.env)) return(set.no.value[2]) } if (!exists(slider.env)) slider.env - new.env() if (!missing(obj.name)) { if (!missing(obj.value)) assign(obj.name, obj.value, env = slider.env) else obj.value - get(obj.name, env = slider.env) return(obj.value) } if (missing(title)) title - slider control widget require(tcltk) nt - tktoplevel() tkwm.title(nt, title) tkwm.geometry(nt, +0+0) if (missing(sl.names)) sl.names - NULL if (missing(sl.functions)) sl.functions - function(...) { } for (i in seq(sl.names)) { eval(parse(text = paste(assign('slider, i, ',tclVar(sl.defaults[i]),env=slider.env), sep = ))) tkpack(fr - tkframe(nt)) lab - tklabel(fr, text = sl.names[i], width = 25) sc - tkscale(fr, from = sl.mins[i], to = sl.maxs[i], showvalue = T, resolution = sl.deltas[i], orient = horiz) tkpack(lab, sc, side = right) assign(sc, sc, env = slider.env) eval(parse(text = paste(tkconfigure(sc,variable=slider, i, ), sep = )), env = slider.env) sl.fun - if (length(sl.functions) 1) sl.functions[[i]] else sl.functions if (!is.function(sl.fun)) sl.fun - eval(parse(text = paste(function(...){, sl.fun, }))) tkconfigure(sc, command = sl.fun) } assign(slider.values.old, sl.defaults, env = slider.env) tkpack(f.but - tkframe(nt), fill = x) tkpack(tkbutton(f.but, text = Exit, command = function() tkdestroy(nt)), side = right) if (missing(reset.function)) reset.function - function(...) print(relax) if (!is.function(reset.function)) reset.function - eval(parse(text = paste(function(...){, reset.function, }))) tkpack(tkbutton(f.but, text = Reset, command = function() { for (i in seq(sl.names)) eval(parse(text = paste(tclvalue(slider, i, )-, sl.defaults[i], sep = )), env = slider.env) reset.function() }), side = right) if (missing(but.names)) but.names - NULL for (i in seq(but.names)) { but.fun - if (length(but.functions) 1) but.functions[[i]] else but.functions if (!is.function(but.fun)) but.fun - eval(parse(text = paste(function(...){, but.fun, }))) tkpack(tkbutton(f.but, text = but.names[i], command = but.fun), side = left) } invisible(nt) } # definition of R.veil.in.the.wind R.veil.in.the.wind-function(){ # Mark Hempelmann / Peter Wolf par(bg=blue4, col=white, col.main=white, col.sub=white, font.sub=2, fg=white) # set colors and fonts samp - function(N,D) N*(1/4+D)/(1/4+D*N) z-outer(seq(1, 800, by=10), seq(.0025, 0.2, .0025)^2/1.96^2, samp) # create 3d matrix h-100 z[10:70,20:25]-z[10:70,20:25]+h; z[65:70,26:45]-z[65:70,26:45]+h z[64:45,43:48]-z[64:45,43:48]+h; z[44:39,26:45]-z[44:39,26:45]+h x-26:59; y-11:38; zz-outer(x,y,+); zz-zz*(65zz)*(zz73) cz-10+col(zz)[zz0];rz-25+row(zz)[zz0]; z[cbind(cz,rz)]-z[cbind(cz,rz)]+h refresh.code-function(...){ theta-slider(no=1); phi-slider(no=2) persp(x=seq(1,800,by=10),y=seq(.0025,0.2,.0025),z=z,theta=theta,phi=phi, scale=T, shade=.9, box=F, ltheta = 45, lphi = 45, col=aquamarine, border=NA,ticktype=detailed) } slider(refresh.code, c(theta, phi), c(0, 0),c(360, 360),c(.2, .2),c(85, 270) ) } # now let's test it! R.veil.in.the.wind() __
Re: [R] Using data frames for EDA: Insert, Change name, delete columns? (Newcomer's question)
Ben Fairbank [EMAIL PROTECTED] wrote in message news:[EMAIL PROTECTED] I ... cannot find commands to easily insert, remove (delete), rename, and re-order (arbitrarily, not sort) columns. ... Could a reader provide a reference where such commands are documented? There's a lot of info in the old R-Help postings, but searching and finding an answer for a particular problem can be a bit of a pain. Here's some info from some old R-Help postings that may help on your question: DELETE TWO COLUMNS --- I have a dataframe 'd2004' and I want to remove two columns: 'd2004$concentration' and 'd2004$stade. I could do it just as follows: names(d2004) [1] Localite Date parcelle maille presence.plant concentration stade.culture [8] stade Trou HorizonProfondeur d2004 - d2004[, -c(6, 8)] but I'd like to use column names (to avoid finding column numbers each time). I cannot find an easy way to operate... I wonder why that works: d2004[, concentration] and this don't: d2004 - d2004[, -c(concentration, stade)] SOLUTIONS: d2004$concentration - NULL d2004$stade - NULL or Newdata - subset(d2004, select=-c(concentration,stade)) RENAMING COLUMNS --- This is a sample data frame: myData - data.frame( col1 = 1:3, col2 = 2:4, col3 = 3:5 ) myData col1 col2 col3 1123 2234 3345 You can change all names by: names( myData )- c( newcol1, newcol2, newcol3 ) myData newcol1 newcol2 newcol3 1 1 2 3 2 2 3 4 3 3 4 5 Or a single name by: names( myData )[ 2 ] - newcol2 myData col1 newcol2 col3 11 23 22 34 33 45 Or if you know the name, but not the column number: names( myData )[ which( names( myData ) == newcol2 ) ] - verynewcol2 myData col1 verynewcol2 col3 11 23 22 34 33 45 REORDERING COLUMNS --- I don't have a clipping for this one, but here's what I'd try: myData - data.frame( col1 = 1:3, col2 = 2:4, col3 = 3:5 ) myData col1 col2 col3 1123 2234 3345 MyData - myData[,c(3,1,2)] MyData col3 col1 col2 1312 2423 3534 -- efg Earl F. Glynn Bioinformatics Stowers Institute for Medical Research __ 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] where can i download the metrics package?
On Tue, 28 Jun 2005 20:55:44 +0800 ronggui wrote: i learn it from metrics: Towards a package for doing econometrics in Rbut i can not find it in cran. I guess you are referring to the talk of Hiroyuki Kawakatsu at useR! 2004. Contact him directly to ask for a version of the package. Some of the functionality contained in the package is by now also provided by some CRAN packages, look at the Econometrics view for further information: http://CRAN.R-project.org/src/contrib/Views/Econometrics.html Z __ 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] where can i download the metrics package?
How about Rmetrics, listed under misc: related projects on www.r-project.org. spencer graves ronggui wrote: i learn it from metrics: Towards a package for doing econometrics in Rbut i can not find it in cran. -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] function for cumulative occurrence of elements
I'm not entirely sure what you want, but is it 9 5 3 for this data? (9 new species occur at the first point, 5 new at the second, and 3 new at the third). If this is right, then to get accumulation curve when random Points are considered, you can probably just index rows of dt appropriately. dd - read.table(clipboard, header=T) dd[,1:3] Pointspecies frequency 1 7 American_elm 7 2 7 apple 2 3 7 black_cherry 8 4 7 black_oak 1 5 7chokecherry 1 6 7 oak_sp 1 7 7 pignut_hickory 1 8 7 red_maple 1 9 7 white_oak 5 10 9 black_spruce 2 11 9blue_spruce 2 12 9missing12 13 9 Norway_spruce 8 14 9 white_spruce 3 1512 apple 2 1612 black_cherry 1 1712 black_locust 1 1812 black_walnut 1 1912 lilac 3 2012missing 2 # dt: table of which species occur at which Points dt - table(dd$Point, dd$species) # doc: for each species, the index of the Point where # it first occurs doc - apply(dt, 2, function(x) which(x==1)[1]) doc American_elm apple black_cherry black_locust black_oak 1 1 1 3 1 black_spruce black_walnutblue_sprucechokecherry lilac 2 3 2 1 3 missing Norway_spruce oak_sp pignut_hickory red_maple 2 2 1 1 1 white_oak white_spruce 1 2 table(doc) doc 1 2 3 9 5 3 hope this helps, Tony Plate Steven K Friedman wrote: Hello, I have a data set with 9700 records, and 7 parameters. The data were collected for a survey of forest communities. Sample plots (1009) and species (139) are included in this data set. I need to determine how species are accumulated as new plots are considered. Basically, I want to develop a species area curve. I've included the first 20 records from the data set. Point represents the plot id. The other parameters are parts of the information statistic H'. Using Table, I can construct a data set that lists the occurrence of a species at any Point (it produces a binary 0/1 data table). From there it get confusing, regarding the most efficient approach to determining the addition of new and or repeated species occurrences. ptcount - table(sppoint.freq$species, sppoint.freq$Point) From here I've played around with colSums to calculate the number of species at each Point. The difficulty is determining if a species is new or repeated. Also since there are 1009 points a function is needed to screen every Point. Two goals are of interest: 1) the species accumulation curve, and 2) an accumulation curve when random Points are considered. Any help would be greatly appreciated. Thank you Steve Friedman Pointspecies frequency point.list point.prop log.prop point.hprime 1 7 American elm 7 27 0.25925926 -1.3499267 0.3499810 2 7 apple 2 27 0.07407407 -2.6026897 0.1927918 3 7 black cherry 8 27 0.29629630 -1.2163953 0.3604134 4 7 black oak 1 27 0.03703704 -3.2958369 0.1220680 5 7chokecherry 1 27 0.03703704 -3.2958369 0.1220680 6 7 oak sp 1 27 0.03703704 -3.2958369 0.1220680 7 7 pignut hickory 1 27 0.03703704 -3.2958369 0.1220680 8 7 red maple 1 27 0.03703704 -3.2958369 0.1220680 9 7 white oak 5 27 0.18518519 -1.6863990 0.3122961 10 9 black spruce 2 27 0.07407407 -2.6026897 0.1927918 11 9blue spruce 2 27 0.07407407 -2.6026897 0.1927918 12 9missing12 27 0. -0.8109302 0.3604134 13 9 Norway spruce 8 27 0.29629630 -1.2163953 0.3604134 14 9 white spruce 3 27 0. -2.1972246 0.2441361 1512 apple 2 27 0.07407407 -2.6026897 0.1927918 1612 black cherry 1 27 0.03703704 -3.2958369 0.1220680 1712 black locust 1 27 0.03703704 -3.2958369 0.1220680 1812 black walnut 1 27 0.03703704 -3.2958369 0.1220680 1912 lilac 3 27 0. -2.1972246 0.2441361 2012missing 2 27 0.07407407 -2.6026897 0.1927918 __ R-help@stat.math.ethz.ch
Re: [R] ks.test() output interpretation
Hi I would recommend graphical methods to compare two samples from possible different distributions. See ?qqplot Since the Kolmogorov-Smirnov test has in many cases very small power, you can not conclude that two sample come from the same distribution only because the ks.test is not significant. The following example shows you one problem: In a short simulation we generate 1000 times two samples (with 100 observation per sample). The first sample has a standard normal distribution, the second a t-distribution with 1 degree of freedom. For each of these 1000 pairs we calculate the ks.test and save the p.value. x1 - matrix(nrow = 100, ncol = 1000) y1 - matrix(nrow = 100, ncol = 1000) test1 - numeric(1000) for(i in 1:1000) { set.seed(i) x1[,i] - rnorm(100) y1[,i] - rt(100, df = 1) test1[i] - ks.test(x1[,i],y1[,i])$p.value } sum(test10.05) Only in 309 of 1000 cases the test shows a significant difference of the two samples. In all other cases we would conclude that the two sample have the same distribution. This is an example with 100 observation per group. If you have smaller groups the power is even worse. If we look at 10 randomly drawn pairs of the 1000 simulations and plot the qqplot: par(mfrow = c(3,3)) ind - sample(1:1000, 9) tmp - sapply(ind, function(j) qqplot(x1[,j],y1[,j], xlab = paste(x1[,,j,]), ylab = paste(y1[,,j,]))) In many cases we see that the two distributions are different. Compare it to the qqplot of two normal distributed random variables: x2 - matrix(rnorm(900), nrow = 100, ncol = 9) y2 - matrix(rnorm(900), nrow = 100, ncol = 9) par(mfrow = c(3,3)) tmp - sapply(1:9, function(j) qqplot(x2[,j],y2[,j], xlab = paste(x2[,,j,]), ylab = paste(y2[,,j,]))) Of course there are situations for which the graphical methods fail, too, but it becomes apparent that it is a descriptive way to describe two distributions. Calculating the Kolmogorov-Smirnov test pretends a clear test result (that two distribution are the same) which is wrong or at least misleading. Best regards, Christoph Buser -- Christoph Buser [EMAIL PROTECTED] Seminar fuer Statistik, LEO C13 ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND phone: x-41-44-632-4673 fax: 632-1228 http://stat.ethz.ch/~buser/ -- kapo coulibaly writes: I'm using ks.test() to compare two different measurement methods. I don't really know how to interpret the output in the absence of critical value table of the D statistic. I guess I could use the p-value when available. But I also get the message cannot compute correct p-values with ties ... does it mean I can't use ks.test() for these data or I can still use the D statistic computed to make a decision whether the two samples come from the same distribution. Thanks!! Rekindle the Rivalries. Sign up for Fantasy Football __ 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-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] How to convert c:\a\b to c:/a/b?
(Ted Harding) wrote: On 27-Jun-05 Spencer Graves wrote: Thanks, Dirk, Gabor, Eric: You all provided appropriate solutions for the stated problem. Sadly, I oversimplified the problem I was trying to solve: I copy a character string giving a DOS path from MS Windows Explorer into an R script file, and I get something like the following: D:\spencerg\statmtds\R\Rnews I want to be able to use this in R with its non-R meaning, e.g., in readLine, count.fields, read.table, etc., after appending a file name. Your three solutions all work for my oversimplified toy example but are inadequate for the problem I really want to solve. The various responses show that solving this problem directly within R may be, well, problematic! I'm not a perl user, but possibly Spencer Graves's speculation might eventually be brought into a straightforward solution. Unfortunately, for this query, I'm not a Windows users either, so can't be authoritative about how practical the following may be for Spencer's problem (and similar). In the Unix world, the string editor program 'sed' simply mops up problems of this kind. For example, I just did: echo D:\spencerg\statmtds\R\Rnews | sed '[EMAIL PROTECTED]@/@g' and got the response: D:/spencerg/statmtds/R/Rnews Note: the parsing of the 'sed' command is as follows: [EMAIL PROTECTED]@[EMAIL PROTECTED] means substitute y for every (g = global) occurrence of x. The character to be replaced (x=\) is the escape character so needs to be escaped (\\); but apart from this it's straightforward. The usual separator is / instead of @, but I used @ to simplify things since the substitute itself is y=/. An alternative using / as separator would be echo D:\spencerg\statmtds\R\Rnews | sed 's/\\/\//g' D:/spencerg/statmtds/R/Rnews which is a bit more complicated but still straightforward (depending on your eyesight). Here, both x=\ and y=/ need to be escaped. 'sed', and a lot of other useful stuff, is available as GNU tools which can be installed on Windows, and allow this kind of thing to be very slickly done, but outside of R of course. I also make much use of 'awk' (but you can equally use perl) for tidying up CSV files exported from Excel, and for all sorts of rearrangements and substitutions in data files. While Unix users take this sort of thing for granted, I suggest to Windows users that it could be well worth installing the GNU tools, since they are very useful indeed. It's not clear from Spencer's follow-up whether doing this kind of preliminary work outside of R, and then bringing the results into R (e.g. via the clipboard or a file) is practical in his context. If not (i.e. all the work has to be done inside R), then of course my sugestion above is not helpful in this case! The closest you can get to the above using R (and on Windows) is echo D:/spencerg/statmtds/R/Rnews tmp.txt echo cat(gsub(, /, readLines(tmp.txt))) | R --slave I do have a question/request to r-devel: In my example, I ideally would like to be able to start an *.R script or alternatively send R expression at the prompt, and then read data from stdin. Two examples: echo D:/spencerg/statmtds/R/Rnews | R --slave --code 'cat(gsub(, /, readLines(tmp.txt)))' or echo cat(gsub(, /, readLines(tmp.txt))) main.R echo D:/spencerg/statmtds/R/Rnews | R --slave --file main.R Note that R CMD BATCH does not provide this. Either you send your R code via standard input or the data, but you cannot send both. If you want to have and dynamic interaction between two applications via stdin and stdout, except from ellaborating with .Rprofile/.First I'm not sure how to start my script in the first place. Henrik Best wishes, Ted. E-Mail: (Ted Harding) [EMAIL PROTECTED] Fax-to-email: +44 (0)870 094 0861 Date: 28-Jun-05 Time: 13:38:37 -- XFMail -- __ 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-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] Producing character given i.e. | with plotmath
Hello! Does someone know how to produce L(y|mu) with plotmath? Some code with unsuccessfull results: plot(dnorm(x = seq(from = -4, to = 4, by = 0.1)), type = l) ## Not what I want legend(legend = c(expression(L(y:mu))), x = topright) ## Strange, is this a bug? legend(legend = c(expression(L(y|mu))), x = top) ## Group produces an error legend(legend = c(expression(group(L(y, |, mu, x = topleft) ## Paste keeps commas in expression legend(legend = c(expression(paste(L(y, |, mu, x = bottomleft) ## This one is OK, but braces are not as they should be legend(legend = c(expression(paste(L(y, |, mu, x = bottom) Thanks! Lep pozdrav / With regards, Gregor Gorjanc -- University of Ljubljana Biotechnical FacultyURI: http://www.bfro.uni-lj.si/MR/ggorjan Zootechnical Department mail: gregor.gorjanc at bfro.uni-lj.si Groblje 3 tel: +386 (0)1 72 17 861 SI-1230 Domzale fax: +386 (0)1 72 17 888 Slovenia, Europe -- One must learn by doing the thing; for though you think you know it, you have no certainty until you try. Sophocles ~ 450 B.C. __ 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] Producing character given i.e. | with plotmath
Gorjanc Gregor wrote: Hello! Does someone know how to produce L(y|mu) with plotmath? Some code with unsuccessfull results: plot(dnorm(x = seq(from = -4, to = 4, by = 0.1)), type = l) ## Not what I want legend(legend = c(expression(L(y:mu))), x = topright) ## Strange, is this a bug? legend(legend = c(expression(L(y|mu))), x = top) No, | is a logical Operator that can be rewritten in its original function form as follows: |(FALSE, TRUE) Hence the result is expected. ## Group produces an error legend(legend = c(expression(group(L(y, |, mu, x = topleft) You have not specified any delimiter. ## Paste keeps commas in expression legend(legend = c(expression(paste(L(y, |, mu, x = bottomleft) correct ## This one is OK, but braces are not as they should be legend(legend = c(expression(paste(L(y, |, mu, x = bottom) What's wrong with the braces?` What you really want is: legend(legend = c(expression(L(group(, y, |) * mu))), x = center) Uwe Ligges Thanks! Lep pozdrav / With regards, Gregor Gorjanc -- University of Ljubljana Biotechnical FacultyURI: http://www.bfro.uni-lj.si/MR/ggorjan Zootechnical Department mail: gregor.gorjanc at bfro.uni-lj.si Groblje 3 tel: +386 (0)1 72 17 861 SI-1230 Domzale fax: +386 (0)1 72 17 888 Slovenia, Europe -- One must learn by doing the thing; for though you think you know it, you have no certainty until you try. Sophocles ~ 450 B.C. __ 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-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] Possible bug in summary of residuals with lm and weights
I sent this to r-devel the other day but didn't get any takers. This may not be a bug but rather an inconsistency. I'm not sure if this is intentional. summary.lm stores weighted residuals whereas I think most users will want print.summary.lm to summarize unweighted ones as if saying summary(resid(fit)). set.seed(1) dat - data.frame(y = rnorm(15), x = rnorm(15), w = 1:15) f - lm(y ~ x, weights = w, data = dat) summary(f) . . . . Residuals: Min 1Q Median 3QMax -8.260 -1.565 0.117 2.105 4.666 . . . . resid(f) 1 2 3 4 5 6 -0.73429677 0.06818092 -1.20558034 1.25783256 0.05231879 -1.18383039 7 8 9 10 11 12 0.16034166 0.59880438 0.98337588 -0.58944957 1.40690588 0.31138819 13 14 15 -0.35111933 -2.20770335 0.89438636 version platform i386-pc-linux-gnu arch i386 os linux-gnu system i386, linux-gnu status major2 minor1.0 year 2005 month04 day 18 language R -- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University __ 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] Possible bug in summary of residuals with lm and weights
On Tue, 28 Jun 2005, Frank E Harrell Jr wrote: I sent this to r-devel the other day but didn't get any takers. This may not be a bug but rather an inconsistency. I'm not sure if this is intentional. summary.lm stores weighted residuals whereas I think most users will want print.summary.lm to summarize unweighted ones as if saying summary(resid(fit)). It seems no one agreed with you! I think most users will want S-compatibility here, and after that residuals that are in some sense on the same scale (that is taking the weights into account). In short, the status quo. I suspect no one wants the definition of the summary.lm class changed. There is a bug in the print method, which has cat(if (!is.null(x$w) diff(range(x$w))) Weighted , Residuals:\n, sep = ) so it is intended to say Weighted Residuals, but w is not in a summary.lm object. set.seed(1) dat - data.frame(y = rnorm(15), x = rnorm(15), w = 1:15) f - lm(y ~ x, weights = w, data = dat) summary(f) . . . . Residuals: Min 1Q Median 3QMax -8.260 -1.565 0.117 2.105 4.666 . . . . resid(f) 1 2 3 4 5 6 -0.73429677 0.06818092 -1.20558034 1.25783256 0.05231879 -1.18383039 7 8 9 10 11 12 0.16034166 0.59880438 0.98337588 -0.58944957 1.40690588 0.31138819 13 14 15 -0.35111933 -2.20770335 0.89438636 version platform i386-pc-linux-gnu arch i386 os linux-gnu system i386, linux-gnu status major2 minor1.0 year 2005 month04 day 18 language R -- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University __ 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] (no subject)
Hello, I was wondering if you can help me I am a final year PhD student and I am trying to do a logistic regression in R and look at the sign of the linear coefficient estimated from this regression in order to decide which model curve I should fit in my data?? How can I do this?? thank you Kelly [[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
Re: [R] factor to character
On Tue, 28 Jun 2005, Omar Lakkis wrote: Using RODBC, when I select from a table strings (chars and varchars) come as factors. What is the best way, speed wise, to convert these columns back to strings (perhaps using as.character). FAQ Q7.10, replacing numeric by character. You do have control in RODBC, in fact: see ?sqlFetch. -- 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] svm and scaling input
Dear All, I've a question about scaling the input variables for an analysis with svm (package e1071). Most of my variables are factors with 4 to 6 levels but there are also some numeric variables. I'm not familiar with the math behind svms, so my assumtions maybe completely wrong ... or obvious. Will the svm automatically expand the factors into a binary matrix? If I add numeric variables outside the range of 0 to 1 do I have to scale them to have 0 to 1 range? thanks a lot for help, +kind regards, Arne __ 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] Using data frames for EDA: Insert, Change name, delete columns? (Newcomer's question)
if i want to rename the data frame,say data1,to data2,have i use those data2-data1 rm(dat1) or these is simpler way to do this? thank you ! On Tue, 28 Jun 2005 10:08:55 -0500 Earl F. Glynn [EMAIL PROTECTED] wrote: Ben Fairbank [EMAIL PROTECTED] wrote in message news:[EMAIL PROTECTED] I ... cannot find commands to easily insert, remove (delete), rename, and re-order (arbitrarily, not sort) columns. ... Could a reader provide a reference where such commands are documented? There's a lot of info in the old R-Help postings, but searching and finding an answer for a particular problem can be a bit of a pain. Here's some info from some old R-Help postings that may help on your question: DELETE TWO COLUMNS --- I have a dataframe 'd2004' and I want to remove two columns: 'd2004$concentration' and 'd2004$stade. I could do it just as follows: names(d2004) [1] Localite Date parcelle maille presence.plant concentration stade.culture [8] stade Trou HorizonProfondeur d2004 - d2004[, -c(6, 8)] but I'd like to use column names (to avoid finding column numbers each time). I cannot find an easy way to operate... I wonder why that works: d2004[, concentration] and this don't: d2004 - d2004[, -c(concentration, stade)] SOLUTIONS: d2004$concentration - NULL d2004$stade - NULL or Newdata - subset(d2004, select=-c(concentration,stade)) RENAMING COLUMNS --- This is a sample data frame: myData - data.frame( col1 = 1:3, col2 = 2:4, col3 = 3:5 ) myData col1 col2 col3 1123 2234 3345 You can change all names by: names( myData )- c( newcol1, newcol2, newcol3 ) myData newcol1 newcol2 newcol3 1 1 2 3 2 2 3 4 3 3 4 5 Or a single name by: names( myData )[ 2 ] - newcol2 myData col1 newcol2 col3 11 23 22 34 33 45 Or if you know the name, but not the column number: names( myData )[ which( names( myData ) == newcol2 ) ] - verynewcol2 myData col1 verynewcol2 col3 11 23 22 34 33 45 REORDERING COLUMNS --- I don't have a clipping for this one, but here's what I'd try: myData - data.frame( col1 = 1:3, col2 = 2:4, col3 = 3:5 ) myData col1 col2 col3 1123 2234 3345 MyData - myData[,c(3,1,2)] MyData col3 col1 col2 1312 2423 3534 -- efg Earl F. Glynn Bioinformatics Stowers Institute for Medical Research __ 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 -- Department of Sociology Fudan University,Shanghai Blog:http://sociology.yculblog.com __ 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] faster algorithm for Kendall's tau
Hi, I need to calculate Kendall's tau for large data vectors (length 100'000). Is somebody aware of a faster algorithm or package function than cor(, method=kendall)? There are ties in the data to be considered (Kendall's tau-b). Any suggestions? Regards Ferdinand __ 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] function for two-part or two-condition models
Two alternatives to the zero inflated Poisson (ZIP) model are mentioned in Jung, Jhun, and Lee (Biometrics, vol 61, no 2, June 2005, p626): Although the ZIP model is more general than the standard Poisson, count data with many zeros are often more dispersed than the ZIP model. In this case, the use of a zero-inflated negative binomial (ZINB) distribution or a zero-inflated generalized Poisson distribution is a good alternative. best, Richard -- Richard E. Remington Statistician KERN Statistical Services, Inc. PO Box 1046 Boise, ID 83701 Tel: 208.426.0113 KernStat.com Andrew Robinson wrote: Hi Richard, I'm not sure that I can imagine how data can have too many zeros to be fit well with zero-inflated Poisson models. Won't the excess zeros be accommodated by increasing the the inflation? In any case, if you want a model that separates the zeros from the occurrences before fitting a Poisson model to account for variation in abundance then it might be safest to do that split manually. Another angle to try is to treat it as a special case of a finite mixture regression. I think that some of Jim Lindsey's code will fit such models. Google can help you find his wbsite. An MS student of mine explored these models for regeneration modeling. I'd be happy to send you a pdf of his thesis if it would help. Cheers, Andrew On Mon, Jun 27, 2005 at 03:35:30PM -0400, Richard Chandler wrote: Hello, This is an (hopefully) improved question of one I posted several weeks ago. Does anyone know of a function for fitting two-part models? These models are designed to handle count data with so many zeroes that they can't be fit well with zero-inflated Poisson models or other 'typical' GLMs. My understanding is that they work by first fitting a binomial model to separate the zeros from the occurrences (positive integers) before fitting a Poisson model to account for variation in abundance. I have tried help.search(two-part) and many other similar guesses. Thanks, Richard -- Richard Chandler, M.S. Candidate Department of Natural Resources Conservation UMass Amherst (413)545-1237 __ 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-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] Circular Mean Question
Hi, a question about the circular mean function in the package CircStats: Can anyone shed some light on why the circ mean function seems to make sense for the first 2 set of bearings and then the mean of 225 and 45 degrees gives an unexpected 180 deg. deg(circ.mean(c(rad(222),rad(45%%360 [1] 133.5 deg(circ.mean(c(rad(224),rad(45%%360 [1] 134.5 deg(circ.mean(c(rad(225),rad(45%%360 [1] 180 deg(circ.mean(c(rad(226),rad(45%%360 [1] 315.5 Can anyone explain this??? This problem was first detected when I was trying to take the circ weighted means of my data: With 2 groups of bearings: x - c(270,180) y - c(45,270) the circular mean of these bearings gives: deg(circ.mean(c(rad(x),rad(y%%360 [1] 257.2356 When finding the weighted means I get this: meany - circ.mean(rad(y)) meanx - circ.mean(rad(x)) deg(circ.weighted.mean(c(meanx,meany),c(2,2)))%%360 [1] 281.25 The function for weighted mean I am using: circ.weighted.mean - function (x,w) { sinr - sum(w*sin(x)) cosr - sum(w*cos(x)) circmean - atan(sinr, cosr) circmean } I am assuming that the problem that mention above is the cause of the different mean bearings. Am I missing something fundamental here? Thanks, Mike __ 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] Brown-Forsythe Test
Hi, do anyone know which statistical software have the Brown-Forsythe Test for differences of numerical continue data groups wich variance heterogenity? not ofr evaluate the homogenity of variances, for evaluated the differences between groups as ANOVA. thanks José Herrera México _ T1msn Search. Todo lo que buscas ahora más rapido http://search.t1msn.com.mx/ __ 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] Brown-Forsythe Test
?oneway.test Jose Herrera Bazán wrote: Hi, do anyone know which statistical software have the Brown-Forsythe Test for differences of numerical continue data groups wich variance heterogenity? not ofr evaluate the homogenity of variances, for evaluated the differences between groups as ANOVA. thanks José Herrera México _ T1msn Search. Todo lo que buscas ahora más rapido http://search.t1msn.com.mx/ __ 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 -- Chuck Cleland, Ph.D. NDRI, Inc. 71 West 23rd Street, 8th floor New York, NY 10010 tel: (212) 845-4495 (Tu, Th) tel: (732) 452-1424 (M, W, F) fax: (917) 438-0894 __ 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] svm and scaling input
[EMAIL PROTECTED] wrote: Dear All, I've a question about scaling the input variables for an analysis with svm (package e1071). Most of my variables are factors with 4 to 6 levels but there are also some numeric variables. I'm not familiar with the math behind svms, so my assumtions maybe completely wrong ... or obvious. Will the svm automatically expand the factors into a binary matrix? If I add numeric variables outside the range of 0 to 1 do I have to scale them to have 0 to 1 range? Well, this depends on the kernel in use. For radial basis functions (as an example), you do not have to rescale, but a transformation of variables might make sense in order to get better results, though. Uwe Ligges thanks a lot for help, +kind regards, Arne __ 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-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] faster algorithm for Kendall's tau
On Tue, 2005-06-28 at 13:03 -0400, ferdinand principia wrote: Hi, I need to calculate Kendall's tau for large data vectors (length 100'000). Is somebody aware of a faster algorithm or package function than cor(, method=kendall)? There are ties in the data to be considered (Kendall's tau-b). Any suggestions? Regards Ferdinand The time intensive part of the process is typically the ranking/ordering of the vector pairs to calculate the numbers of concordant and discordant pairs. If the number of _unique pairs_ in your data is substantially less than the number of total pairs (in other words, creating a smaller 2d contingency table from a pair of your vectors makes sense), then the following may be of help. # Calculate CONcordant Pairs in a table # cycle through x[r, c] and multiply by # sum(x elements below and to the right of x[r, c]) # x = table concordant - function(x) { x - matrix(as.numeric(x), dim(x)) # get sum(matrix values r AND c) # for each matrix[r, c] mat.lr - function(r, c) { lr - x[(r.x r) (c.x c)] sum(lr) } # get row and column index for each # matrix element r.x - row(x) c.x - col(x) # return the sum of each matrix[r, c] * sums # using mapply to sequence thru each matrix[r, c] sum(x * mapply(mat.lr, r = r.x, c = c.x)) } # Calculate DIScordant Pairs in a table # cycle through x[r, c] and multiply by # sum(x elements below and to the left of x[r, c]) # x = table discordant - function(x) { x - matrix(as.numeric(x), dim(x)) # get sum(matrix values r AND c) # for each matrix[r, c] mat.ll - function(r, c) { ll - x[(r.x r) (c.x c)] sum(ll) } # get row and column index for each # matrix element r.x - row(x) c.x - col(x) # return the sum of each matrix[r, c] * sums # using mapply to sequence thru each matrix[r, c] sum(x * mapply(mat.ll, r = r.x, c = c.x)) } # Calculate Kendall's Tau-b # x = table calc.KTb - function(x) { x - matrix(as.numeric(x), dim(x)) c - concordant(x) d - discordant(x) n - sum(x) SumR - rowSums(x) SumC - colSums(x) KTb - (2 * (c - d)) / sqrt(((n ^ 2) - (sum(SumR ^ 2))) * ((n ^ 2) - (sum(SumC ^ 2 KTb } Note that I made some modifications of the above, relative to prior versions that I have posted to handle large numbers of pairs to avoid integer overflows in summations. Hence the: x - matrix(as.numeric(x), dim(x)) conversion in each function. Now, create some random test data, with 100,000 elements in each vector, sampling from 'letters', which would yield a 26 x 26 table: a - sample(letters, 10, replace = TRUE) b - sample(letters, 10, replace = TRUE) dim(table(a, b)) [1] 26 26 system.time(print(calc.KTb(table(a, b [1] 0.0006906088 [1] 0.77 0.02 0.83 0.00 0.00 Note that in the above, the initial table takes most of the time: system.time(table(a, b)) [1] 0.55 0.00 0.56 0.00 0.00 Hence: tab.ab - table(a, b) system.time(print(calc.KTb(tab.ab))) [1] 0.0006906088 [1] 0.25 0.01 0.27 0.00 0.00 I should note that I also ran: system.time(print(cor(a, b, method = kendall))) [1] 0.0006906088 [1] 694.80 7.72 931.89 0.00 0.00 Nice to know the results work out at least... :-) I have not tested with substantially larger 2d matrices, but would envision that as the dimensions of the resultant tabulation increases, my method probably approaches and may even become less efficient than the approach implemented in cor(). Some testing would validate this and perhaps point to coding the concordant() and discordant() functions in C for improvement in timing. HTH, Marc Schwartz __ 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] 3D ellipsoid confidence region
I am curious if there is code developed to plot confidence regions in 3D. The scatterplot3d function generates the plot I want, but would like an 3D equivalent to the data.ellipse function. Any help in this direction would be appreciated, be it theoretical, graphical, or otherwise. Melanie Edwards Senior Statistician Exponent 15375 SE 30th PL, Suite 250 Bellevue, WA 98007 Tel: (425) 519-8714 Fax: (425) 643-9827 Cell: (206) 852-5739 www.exponent.com __ 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] Course***R/S-plus Fundamentals and Programming Techinques: Seattle July 11-12
XLSolutions Corporation (www.xlsolutions-corp.com) is proud to announce 2-day R/S-plus Fundamentals and Programming Techniques at 4 locations nationwide. www.xlsolutions-corp.com/Rfund.htm Seattle, WA - July 11th-12th, 2005 Reserve your seat now at the early bird rates! Payment due AFTER the class Course Description: This two-day beginner to intermediate R/S-plus course focuses on a broad spectrum of topics, from reading raw data to a comparison of R and S. We will learn the essentials of data manipulation, graphical visualization and R/S-plus programming. We will explore statistical data analysis tools,including graphics with data sets. How to enhance your plots, build your own packages (librairies) and connect via ODBC,etc. We will perform some statistical modeling and fit linear regression models. Participants are encouraged to bring data for interactive sessions With the following outline: - An Overview of R and S - Data Manipulation and Graphics - Using Lattice Graphics - A Comparison of R and S-Plus - How can R Complement SAS? - Writing Functions - Avoiding Loops - Vectorization - Statistical Modeling - Project Management - Techniques for Effective use of R and S - Enhancing Plots - Using High-level Plotting Functions - Building and Distributing Packages (libraries) - Connecting; ODBC, Rweb, Orca via sockets and via Rjava Interested in R/Splus Advanced course? email us. Email us for group discounts. Email Sue Turner: [EMAIL PROTECTED] Phone: 206-686-1578 Visit us: www.xlsolutions-corp.com/training.htm Please let us know if you and your colleagues are interested in this classto take advantage of group discount. Register now to secure your seat! Interested in R/Splus Advanced course? email us. Cheers, Elvis Miller, PhD Manager Training. XLSolutions Corporation 206 686 1578 www.xlsolutions-corp.com [EMAIL PROTECTED] __ 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] Circular Mean Question
Hi Mike, You should think of vector sums rather than arithmetic sum. So the mean is really the direction of the resultant vector of all the unit vectors. When two unit vectors are exactly opposing each other (as in your example with angles 45 and 225 degrees), they cancel each other out and the resultant vector has length zero, and hence the direction is arbitrary and undefined. However, there are two issues which give rise to seemingly anomalous results: (1) due to finite numerical precision in computing the sines and cosines, exact cancellation does not happen, and (2) discontinuity in atan function due to branch-point (0, 0) and a branch cut along negative real axis. For example, deg(circ.mean(c(rad(30),rad(210 [1] -56.30993247402022 sx - sin(rad(210)) + sin(rad(30)) cx - cos(rad(210)) + cos(rad(30)) sx [1] -1.665334536937735e-16 cx [1] 1.110223024625157e-16 atan(sx,cx) # this should really be 0 [1] -0.9827937232473291 If you now look at the resultant vector it has components (sin(-0.9828), cos(-0.9828)) = (-0.83,0.55), which is completely wrong. However, the above example is only a minor perturbation away from the following (where the resultant vector is not zero): deg(circ.mean(c(rad(211),rad(30 # note 211 = -149 (mod 360) [1] -59.46 which is perfectly alright. Hope this helps, Ravi. -- Ravi Varadhan, Ph.D. Assistant Professor, The Center on Aging and Health Division of Geriatric Medicine and Gerontology Johns Hopkins University Ph: (410) 502-2619 Fax: (410) 614-9625 Email: [EMAIL PROTECTED] -- -Original Message- From: [EMAIL PROTECTED] [mailto:r-help- [EMAIL PROTECTED] On Behalf Of Michael Peckford Sent: Tuesday, June 28, 2005 1:35 PM To: r-help@stat.math.ethz.ch Subject: [R] Circular Mean Question Hi, a question about the circular mean function in the package CircStats: Can anyone shed some light on why the circ mean function seems to make sense for the first 2 set of bearings and then the mean of 225 and 45 degrees gives an unexpected 180 deg. deg(circ.mean(c(rad(222),rad(45%%360 [1] 133.5 deg(circ.mean(c(rad(224),rad(45%%360 [1] 134.5 deg(circ.mean(c(rad(225),rad(45%%360 [1] 180 deg(circ.mean(c(rad(226),rad(45%%360 [1] 315.5 Can anyone explain this??? This problem was first detected when I was trying to take the circ weighted means of my data: With 2 groups of bearings: x - c(270,180) y - c(45,270) the circular mean of these bearings gives: deg(circ.mean(c(rad(x),rad(y%%360 [1] 257.2356 When finding the weighted means I get this: meany - circ.mean(rad(y)) meanx - circ.mean(rad(x)) deg(circ.weighted.mean(c(meanx,meany),c(2,2)))%%360 [1] 281.25 The function for weighted mean I am using: circ.weighted.mean - function (x,w) { sinr - sum(w*sin(x)) cosr - sum(w*cos(x)) circmean - atan(sinr, cosr) circmean } I am assuming that the problem that mention above is the cause of the different mean bearings. Am I missing something fundamental here? Thanks, Mike __ 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-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] lattice graph background
Hi user's How to modify the color background on lattice plot. I am trying to plotting histogram from lattice package. Thank very much * Juan Carlos Quiroz Instituto de Fomento Pesquero CHILE __ 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] lattice graph background
Juan Carlos Quiroz Espinosa wrote: Hi user's How to modify the color background on lattice plot. I am trying to plotting histogram from lattice package. Thank very much Try: trellis.par.set(theme = col.whitebg()) or create your own theme. See ?trellis.par.set. HTH, --sundar __ 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] How to extract the within group correlation structure matrix in lme
Dear R users, I fitted a repeated measure model without random effects by using lme. I will use the estimates from that model as an initial estimates to do multiple imputation for missing values of the response variable in the model. I am trying to extract the within group correlation matrix or covariance matrix. here is my code: f = lme(y ~x0+x1+trt+tim+x1:tim +tim:trt,random=~-1|subj, data=dat,corr =corAR1()) f$sigma [1] 2.330854 b=summary(f)$modelStruct$corStruct b Correlation structure of class corAR1 representing Phi 0.8518711 I think 0.8518711 and f$sigma is what I need to reconstruct the variance-cavariance matrix. So, How can I extract 0.8518711 so that I can assign it to a variable? Also, I don't understand what the following parameters estimates mean? summary(f)$modelStruct reStruct parameters: subj -20.15833 corStruct parameters: [1] 2.525869 I appreciate your time on this. Best wishes, Liqiu - Rekindle the Rivalries. Sign up for Fantasy Football [[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
Re: [R] 3D ellipsoid confidence region
Dear Melanie, I think that you're referring to the data.ellipse function in the car package. What you're suggesting is a nice idea but I haven't done it. I think that it might be easier to draw the ellipsoid using the rgl package rather than scatterplot3d. (See, e.g., the scatter3d function in the Rcmdr package.) Regards, John On Tue, 28 Jun 2005 12:35:35 -0700 Melanie Edwards [EMAIL PROTECTED] wrote: I am curious if there is code developed to plot confidence regions in 3D. The scatterplot3d function generates the plot I want, but would like an 3D equivalent to the data.ellipse function. Any help in this direction would be appreciated, be it theoretical, graphical, or otherwise. Melanie Edwards Senior Statistician Exponent 15375 SE 30th PL, Suite 250 Bellevue, WA 98007 Tel: (425) 519-8714 Fax: (425) 643-9827 Cell: (206) 852-5739 www.exponent.com __ 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 John Fox Department of Sociology McMaster University Hamilton, Ontario, Canada http://socserv.mcmaster.ca/jfox/ __ 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] How to extract the within group correlation structure matrix in lme
Have you tried VarCorr(f)? spencer graves liqiu jiang wrote: Dear R users, I fitted a repeated measure model without random effects by using lme. I will use the estimates from that model as an initial estimates to do multiple imputation for missing values of the response variable in the model. I am trying to extract the within group correlation matrix or covariance matrix. here is my code: f = lme(y ~x0+x1+trt+tim+x1:tim +tim:trt,random=~-1|subj, data=dat,corr =corAR1()) f$sigma [1] 2.330854 b=summary(f)$modelStruct$corStruct b Correlation structure of class corAR1 representing Phi 0.8518711 I think 0.8518711 and f$sigma is what I need to reconstruct the variance-cavariance matrix. So, How can I extract 0.8518711 so that I can assign it to a variable? Also, I don't understand what the following parameters estimates mean? summary(f)$modelStruct reStruct parameters: subj -20.15833 corStruct parameters: [1] 2.525869 I appreciate your time on this. Best wishes, Liqiu - Rekindle the Rivalries. Sign up for Fantasy Football [[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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] JGR font rendering on Windows
JGR users on Windows may be interested that Sun has released a preview of its latest Java development kit and runtime engine at http://www.java.net/download/jdk6/binaries/ This removes the main problem I have had with JGR and other Java programs on Windows: poor font rendering. On Windows XP with an LCD display, the improvement is striking, as this uses Windows native font smoothing (CoolType). I believe it uses the default Windows font smoothing on other versions of Windows or other displays. MHP -- Michael Prager, Ph.D. NOAA Center for Coastal Fisheries and Habitat Research Beaufort, North Carolina 28516 http://shrimp.ccfhrb.noaa.gov/~mprager/ *** Opinions are personal, not official. No government endorsement is made of any commercial or noncommercial product. *** __ 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] conversion
Dear List, How can I convert a list with elements being character strings, like: c(1,2,3,4), “c(1,3,4,2) … to a list with elements as numerical vectors: c(1,2,3,4), c(1,3,4,2)…? Thanks! Anna __ 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] Reference Card (especially useful for Newbies)
Newbies (and others!) may find useful the R Reference Card made available by Tom Short and Rpad at http://www.rpad.org/Rpad/Rpad-refcard.pdf or through the Contributed link on CRAN (where some other reference cards are also linked). It categorizes and organizes a bunch of R's basic, most used functions so that they can be easily found. For example, paste() is under the Strings heading and expand.grid() is under Data Creation. For newbies struggling to find the right R function as well as veterans who can't quite remember the function name, it's very handy. -- Bert Gunter Genentech Non-Clinical Statistics South San Francisco, CA The business of the statistician is to catalyze the scientific learning process. - George E. P. Box __ 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-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] test of hazard ratios
Can you fit a single model that includes the two classifiers from which the model with one classifier could be obtained as a special case? If yes (and if neither parameter was at a boundary crudely similar to a zero variance component in lme), then you could use the theory that 2*ln(likelihood ratio) is approximately chi-square. For many capabilities in R, this is implemented as a function anova(fit2, fit1) where fit2 and fit1 are the results of fitting the general and the specialized models. If you don't see how to make this work, PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html;. It might help you find a better answer. Failing that, if you submit another post after following the process described therein, it should increase the chances that you will get a useful reply. spencer graves Steve Adams wrote: Hi, Is there a R test available that tests whether 2 hazard ratios obtained from Cox regressions on the same patient sample by 2 different classifiers are significantly different? Thanks Steve __ 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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] Help with stripplot
For the following code is there a way to make the jitter all line up horizontally, instead of them being just randomly spread around a value. So for ex if there are multiple values at 63 for genotype wt then all the values should be plotted on the same y value of 63 but spaced apart by a certain factor or noise.. library(lattice); dataFrame - as.data.frame(t(structure(c( 64,'wt', 62,'wt', 66,'wt', 65,'wt', 60,'wt', 61,'wt', 65,'wt', 66,'wt', 65,'wt', 63,'wt', 67,'wt', 65,'wt', 62,'wt', 65,'wt', 68,'wt', 65,'wt', 63,'wt', 65,'wt', 62,'wt', 65,'wt', 66,'wt', 62,'wt', 65,'wt', 65,'wt', 66,'wt', 65,'wt', 62,'wt', 65,'wt', 66,'wt', 65,'wt', 61,'wt', 65,'wt', 66,'wt', 65,'wt', 62,'wt', 63,'het', 67,'het', 60,'het', 67,'het', 66,'het', 62,'het', 65,'het', 62,'het', 61,'het', 62,'het', 66,'het', 60,'het', 65,'het', 65,'het', 61,'het', 64,'het', 68,'het', 64,'het', 63,'het', 62,'het', 64,'het', 62,'het', 64,'het', 65,'het', 60,'het', 65,'het', 70,'het', 63,'het', 67,'het', 66,'het', 65,'hom', 62,'hom', 68,'hom', 67,'hom', 67,'hom', 63,'hom', 67,'hom', 66,'hom', 63,'hom', 72,'hom', 62,'hom', 61,'hom', 66,'hom', 64,'hom', 60,'hom', 61,'hom', 66,'hom', 66,'hom', 66,'hom', 62,'hom', 70,'hom', 65,'hom', 64,'hom', 63,'hom', 65,'hom', 69,'hom', 61,'hom', 66,'hom', 65,'hom', 61,'hom', 63,'hom', 64,'hom', 67,'hom'), .Dim=c(2,98; colnames(dataFrame) - c('marbles_buried', 'genotype'); dataFrame[c(marbles_buried)] - lapply(dataFrame[c(marbles_buried)], function(x) as.numeric(levels(x)[x])); trellis.par.set(theme = col.whitebg()); stripplot(jitter(marbles_buried) ~ genotype, data = dataFrame, aspect = 1, jitter = TRUE, xlab='Genotype', ylab = Marbles Buried, main='MBA WTs Vs HOMs'); Thanks, Sandeep Kumar Ghosh Lexicon Genetics, Inc., 8800 Technology Forest, The Woodlands, TX 77380 (281) 863-3479 Internal Extension - 3479 [[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
Re: [R] conversion
Anna Oganyan wrote: Dear List, How can I convert a list with elements being character strings, like: c(1,2,3,4), “c(1,3,4,2) … to a list with elements as numerical vectors: c(1,2,3,4), c(1,3,4,2)…? Thanks! Anna Try: x - list(c(1,2,3,4), c(1,3,4,2)) lapply(x, function(x) eval(parse(text = x))) HTH, --sundar __ 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] sample R code for multiple imputation
Hi, I have a big dataset which has many missing values and want to implement Multiple imputation via Monte carlo markov chain by following J Schafer's Analysis of incomplete multivariate data. I don't know where to begin and is looking for a sample R code that implements multiple imputation with EM, MCMC, etc Any help / suggestion will be greatly appreciated. David __ 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] conversion
On Tue, 2005-06-28 at 17:23 -0400, Anna Oganyan wrote: Dear List, How can I convert a list with elements being character strings, like: c(1,2,3,4), “c(1,3,4,2) … to a list with elements as numerical vectors: c(1,2,3,4), c(1,3,4,2)…? Thanks! Anna l - list(c(1,2,3,4), c(1,3,4,2)) l [[1]] [1] c(1,2,3,4) [[2]] [1] c(1,3,4,2) Now use lapply() over each list element in 'l', converting the character vectors to R expressions and then evaluating them: lapply(l, function(x) eval(parse(text = x))) [[1]] [1] 1 2 3 4 [[2]] [1] 1 3 4 2 See ?lapply, ?eval and ?parse. HTH, Marc Schwartz __ 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] How to sort a dataframe by one variable
Hi there, Could anybody help me on how to sort a dataframe by one variable in the dataframe? Thank you Lisa Wang Princess Margaret Hospital Toronto, Ca tel: 416 946 4501 __ 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] conversion
as.numeric? spencer graves Anna Oganyan wrote: Dear List, How can I convert a list with elements being character strings, like: c(1,2,3,4), “c(1,3,4,2) … to a list with elements as numerical vectors: c(1,2,3,4), c(1,3,4,2)…? Thanks! Anna __ 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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] Help with stripplot
Ghosh, Sandeep wrote: For the following code is there a way to make the jitter all line up horizontally, instead of them being just randomly spread around a value. So for ex if there are multiple values at 63 for genotype wt then all the values should be plotted on the same y value of 63 but spaced apart by a certain factor or noise.. snip Yes, remove the jitter function call: stripplot(marbles_buried ~ genotype, data = dataFrame, aspect = 1, jitter = TRUE, xlab='Genotype', ylab = Marbles Buried, main='MBA WTs Vs HOMs') --sundar __ 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] R-help Digest, Vol 28, Issue 28
On Tuesday 28 June 2005 15:30, [EMAIL PROTECTED] wrote: Re : 37. Re: A. Mani : colours in Silhouette (Mulholland, Tom) Message: 37 Date: Tue, 28 Jun 2005 09:08:24 +0800 From: Mulholland, Tom [EMAIL PROTECTED] Subject: Re: [R] A. Mani : colours in Silhouette To: [EMAIL PROTECTED], r-help@stat.math.ethz.ch Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=iso-8859-1 It's not so much a problem, as not working the way you expected. cluster:::plot.partition is used to do the plotting. If you look at the code for this you can see the difficulty in putting every possible permutation into the code. If for example you want the silhouette plot to be red using col = red is not intuitive as the cluster plot (which comes up first) has more than one colour. If you have a look at methods(plot) (assuming that you have loaded the cluster package) you will see that there is a specific piece of code in the form of plot.silhouette. It has an asterisk next to it so you need to use cluster:::plot.silhouette to see the code. It has what you need. args(cluster:::plot.silhouette) function (x, nmax.lab = 40, max.strlen = 5, main = NULL, sub = NULL, xlab = expression(Silhouette width * s[i]), col = gray, do.col.sort = length(col) 1, border = 0, cex.names = par(cex.axis), do.n.k = TRUE, do.clus.stat = TRUE, ...) data(ruspini) pr4 - pam(ruspini, 4) si - silhouette(pr4) plot(si,col = red) I tried that before with many more options and got a blank image. It must have been due to the options. The issue is that whenever code is written there is always a choice as to what functionality is put in place. Just because something can be done, does not mean it will or in some cases should be done. In this case the help for plot.partition notes that For more flexibility, use 'plot(silhouette(x), ...)', see 'plot.silhouette'. Tom Thanks for that I found out something I will find useful in the future. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of A. Mani Sent: Tuesday, 28 June 2005 4:30 AM To: r-help@stat.math.ethz.ch Subject: [R] A. Mani : colours in Silhouette Hello, In cluster analysis with cluster, how does one colour the silhouette plots ? For example in using pam. There seems to be some problem there. Everything else can be coloured. Thanks, A. Mani Member, Cal. Math. Soc __ 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] How to sort a dataframe by one variable
Lisa Wang wrote: Hi there, Could anybody help me on how to sort a dataframe by one variable in the dataframe? Thank you See ?order. x - data.frame(a = runif(10), b = 1:10) x[order(x$a), ] --sundar __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] How to sort a dataframe by one variable
Try RSiteSearch(Sort data frame). Anyhow, use order() or sort.list() as follows: dat[order(dat[,'myColumn']),]#Sorts data.frame dat by column myColumn dat-dat[sort.list(dat$myColumn, decreasing = T),] #Another option to do the same. You can specify decreasing or increasing within both function Cheers Francisco From: Lisa Wang [EMAIL PROTECTED] To: R-Help r-help@stat.math.ethz.ch Subject: [R] How to sort a dataframe by one variable Date: Tue, 28 Jun 2005 17:39:40 -0400 Hi there, Could anybody help me on how to sort a dataframe by one variable in the dataframe? Thank you Lisa Wang Princess Margaret Hospital Toronto, Ca tel: 416 946 4501 __ 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-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] faster algorithm for Kendall's tau
Sorry, I should specifiy in more detail what my data looks like. The data vectors (simulations) are mostly composed of floats (for which it's pretty unlikely to produce ties), but there are integer values to be found as well (up to 10% of vector elements). As I undestand, Marc's algo is not suited for this case. Is there another solution? Thanks Ferdinand --- Marc Schwartz (via MN) [EMAIL PROTECTED] wrote: On Tue, 2005-06-28 at 13:03 -0400, ferdinand principia wrote: Hi, I need to calculate Kendall's tau for large data vectors (length 100'000). Is somebody aware of a faster algorithm or package function than cor(, method=kendall)? There are ties in the data to be considered (Kendall's tau-b). Any suggestions? Regards Ferdinand The time intensive part of the process is typically the ranking/ordering of the vector pairs to calculate the numbers of concordant and discordant pairs. If the number of _unique pairs_ in your data is substantially less than the number of total pairs (in other words, creating a smaller 2d contingency table from a pair of your vectors makes sense), then the following may be of help. # Calculate CONcordant Pairs in a table # cycle through x[r, c] and multiply by # sum(x elements below and to the right of x[r, c]) # x = table concordant - function(x) { x - matrix(as.numeric(x), dim(x)) # get sum(matrix values r AND c) # for each matrix[r, c] mat.lr - function(r, c) { lr - x[(r.x r) (c.x c)] sum(lr) } # get row and column index for each # matrix element r.x - row(x) c.x - col(x) # return the sum of each matrix[r, c] * sums # using mapply to sequence thru each matrix[r, c] sum(x * mapply(mat.lr, r = r.x, c = c.x)) } # Calculate DIScordant Pairs in a table # cycle through x[r, c] and multiply by # sum(x elements below and to the left of x[r, c]) # x = table discordant - function(x) { x - matrix(as.numeric(x), dim(x)) # get sum(matrix values r AND c) # for each matrix[r, c] mat.ll - function(r, c) { ll - x[(r.x r) (c.x c)] sum(ll) } # get row and column index for each # matrix element r.x - row(x) c.x - col(x) # return the sum of each matrix[r, c] * sums # using mapply to sequence thru each matrix[r, c] sum(x * mapply(mat.ll, r = r.x, c = c.x)) } # Calculate Kendall's Tau-b # x = table calc.KTb - function(x) { x - matrix(as.numeric(x), dim(x)) c - concordant(x) d - discordant(x) n - sum(x) SumR - rowSums(x) SumC - colSums(x) KTb - (2 * (c - d)) / sqrt(((n ^ 2) - (sum(SumR ^ 2))) * ((n ^ 2) - (sum(SumC ^ 2 KTb } Note that I made some modifications of the above, relative to prior versions that I have posted to handle large numbers of pairs to avoid integer overflows in summations. Hence the: x - matrix(as.numeric(x), dim(x)) conversion in each function. Now, create some random test data, with 100,000 elements in each vector, sampling from 'letters', which would yield a 26 x 26 table: a - sample(letters, 10, replace = TRUE) b - sample(letters, 10, replace = TRUE) dim(table(a, b)) [1] 26 26 system.time(print(calc.KTb(table(a, b [1] 0.0006906088 [1] 0.77 0.02 0.83 0.00 0.00 Note that in the above, the initial table takes most of the time: system.time(table(a, b)) [1] 0.55 0.00 0.56 0.00 0.00 Hence: tab.ab - table(a, b) system.time(print(calc.KTb(tab.ab))) [1] 0.0006906088 [1] 0.25 0.01 0.27 0.00 0.00 I should note that I also ran: system.time(print(cor(a, b, method = kendall))) [1] 0.0006906088 [1] 694.80 7.72 931.89 0.00 0.00 Nice to know the results work out at least... :-) I have not tested with substantially larger 2d matrices, but would envision that as the dimensions of the resultant tabulation increases, my method probably approaches and may even become less efficient than the approach implemented in cor(). Some testing would validate this and perhaps point to coding the concordant() and discordant() functions in C for improvement in timing. HTH, Marc Schwartz __ 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] the function of mle
I'd guess Jun meant the mle() in the stats4 package. It comes with R, but is not loaded by default, so one would need to do library(stats4) first. Andy From: Spencer Graves Did you do install.packages('nlme') and then 'library(nlme)' before trying to use 'lme'? spencer graves Liu Jun wrote: Hi, I am a user of R computer language.Several days ago,I download R-2.1.0.From the help manual,I saw there is a function mle to find the maximum likelihood estimators.But when I try to use this function,there is no this function.Could you help me to solve this problem? I am looking forward to your answer. Thank you. Jun Liu _ 免费下载 MSN Explorer: http://explorer.msn.com/lccn/ -- -- __ 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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] setting yranges of boxplots
Hi all, I was wondering how to create a graph (boxplot) setting the y range of the graph manually. Something like this: x - c(1:100) boxplot(x, yrange(0, 1000)) Thanks for your time, Adrian __ 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] setting yranges of boxplots
Something like this? boxplot(runif(100, 0, 100), ylim=c(0, 1000)) Andy From: Adrian L. Garcia-Lomana Hi all, I was wondering how to create a graph (boxplot) setting the y range of the graph manually. Something like this: x - c(1:100) boxplot(x, yrange(0, 1000)) Thanks for your time, Adrian __ 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-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] looking for the source
I'm working with the kmeans function in the stats package. This function calls complied fortran program: .Fortran(kmns, as.double(x).) Is there a way to get the source code for this compiled program kmns?? Philip Bermingham [[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] Random Forests theoretical question
I feel like a dope for asking this, but after reading various articles I'm still not quite sure I get it. What is the relationship or similarity between a probability prediction from a classfication random forest and a Bayesian posterior probability? Can they be considered similar, or would I need to take that probability and project it through a likelihood (I'm thinking, perhaps wrongly, of computing this from the confusion matrix) to get a posterior probability? Something about that feels wrong, and not being a real statistician (Just a shrink with some applied stat training) I can't quite connect the dots. Before, this, I've been playing with RF as a way to get a feel for what's going on in the data, rather than for making any real-world predictions. I just don't want to over-interpret (or misinterpret) those probabilities. From Breiman's site, this is probably clear to a real statistician, but someone like me needs to be clobbered over the head with it :-/. Wisdom Welcome! Humbly, DVB [[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] quick way to construct formula
Dear R users, I have a data with 1000 variables named x1, x2, ..., x1000, and I want to construct a formula like this format: ~x1+x2+...+x1000+x1:x2+x1:x3+x999:x1000+log(x1)+...+log(x1000) That is: the base variables followed by all interaction terms and all base feature log-transformations. I know I can use several paste functions to construct it. But is there any other handy way to do it? -Luke __ 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] looking for the source
Off course. That's the beauty of open source :) You will find the source code at http://cran.r-project.org/ under R Sources. Cheers Francisco From: Philip Bermingham [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Subject: [R] looking for the source Date: Tue, 28 Jun 2005 22:36:04 -0400 I'm working with the kmeans function in the stats package. This function calls complied fortran program: .Fortran(kmns, as.double(x).) Is there a way to get the source code for this compiled program kmns?? Philip Bermingham [[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-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] How to convert c:\a\b to c:/a/b
I couldn't resist adding a more literal answer unback - function(x) { chars - unlist(strsplit(deparse(x),)) chars - chars[-c(1,length(chars))] paste(gsub(,/,chars),collapse=) } unback(\n) | David Duffy (MBBS PhD) ,-_|\ | email: [EMAIL PROTECTED] ph: INT+61+7+3362-0217 fax: -0101 / * | Epidemiology Unit, Queensland Institute of Medical Research \_,-._/ | 300 Herston Rd, Brisbane, Queensland 4029, Australia GPG 4D0B994A v __ 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] How to convert c:\a\b to c:/a/b
Close, but does not work generally with RGui 2.1.1 patched, Windows XP: unback(x=n\o) [1] no I'm also unable to parse echo, suggested by Ten Harding and Henrik Bengtsson: echo D:/spencerg/statmtds/R/Rnews tmp.txt Error: syntax error echo cat(gsub(, /, readLines(tmp.txt))) | R --slave Error: syntax error Earlier today, Sundar Dorai-Raj helped me with the following: (File0 - file.choose()) [1] D:\\spencerg\\dataPOWER\\stats\\Tukey\\Boxplot_missing_Tukey2.txt strsplit(File0, ) [[1]] [1] D: spencerg [3] dataPOWER stats [5] Tukey Boxplot_missing_Tukey2.txt fp. - strsplit(File0, )[[1]] (path - paste(fp.[-length(fp.)], collapse=/)) [1] D:/spencerg/dataPOWER/stats/Tukey setwd(path) getwd() [1] D:/spencerg/dataPOWER/stats/Tukey File - fp.[length(fp.)] File [1] Boxplot_missing_Tukey2.txt Thanks to everyone who has contributed (or even read) this thread. I'm confident that a better method exists. Best Wishes, Spencer Graves David Duffy wrote: I couldn't resist adding a more literal answer unback - function(x) { chars - unlist(strsplit(deparse(x),)) chars - chars[-c(1,length(chars))] paste(gsub(,/,chars),collapse=) } unback(\n) | David Duffy (MBBS PhD) ,-_|\ | email: [EMAIL PROTECTED] ph: INT+61+7+3362-0217 fax: -0101 / * | Epidemiology Unit, Queensland Institute of Medical Research \_,-._/ | 300 Herston Rd, Brisbane, Queensland 4029, Australia GPG 4D0B994A v __ 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 -- Spencer Graves, PhD Senior Development Engineer PDF Solutions, Inc. 333 West San Carlos Street Suite 700 San Jose, CA 95110, USA [EMAIL PROTECTED] www.pdf.com http://www.pdf.com Tel: 408-938-4420 Fax: 408-280-7915 __ 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] sample R code for multiple imputation
There is also the mice package at http://www.multiple-imputation.com/ which uses mcmc but is different to Schafer's packages. Simon. At 09:30 AM 29/06/2005, James Reilly wrote: Schafer's functions have been ported to R in the packages norm, cat, mix and pan. Their documentation includes sample code illustrating how to use them. The aregImpute function in Hmisc provides a range of other imputation models, if you're not set on using MCMC. The mitools package might also be useful for dealing with the imputed values. Hope this helps, James On 29/06/2005 9:32 a.m., David Hwang wrote: Hi, I have a big dataset which has many missing values and want to implement Multiple imputation via Monte carlo markov chain by following J Schafer's Analysis of incomplete multivariate data. I don't know where to begin and is looking for a sample R code that implements multiple imputation with EM, MCMC, etc Any help / suggestion will be greatly appreciated. David Simon Blomberg, B.Sc.(Hons.), Ph.D, M.App.Stat. Centre for Resource and Environmental Studies The Australian National University Canberra ACT 0200 Australia T: +61 2 6125 7800 email: Simon.Blomberg_at_anu.edu.au F: +61 2 6125 0757 CRICOS Provider # 00120C __ 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