[R] lme output
Dear all, I noticed the following in the call of lme using msVerbose. fm1 - lme(distance ~ age, data = Orthodont, control = lmeControl(msVerbose=T)) 9 318.073: -0.567886 0.152479 1.98021 10 318.073: -0.567191 0.152472 1.98009 11 318.073: -0.567208 0.152473 1.98010 fm2 - lme(distance ~ age, random =~age, data = Orthodont, lmeControl(msVerbose=T)) 7 318.073: -0.342484 1.75530 4.44650 8 318.073: -0.342507 1.75539 4.44614 9 318.073: -0.342497 1.75539 4.44614 The two model are equivalent and give the same estimates. However, the optimal parameters in the profiled log-likelihood are not the same? why? As I usually thought, the parameters optimised in the profiled likelihood are the log of the precision matrix. The latter can be derived as a Cholesky factorization of the product between the residuals variance and the inverse of the random effects covariance. When I check that it's not the case for model fm1 even if it's equivalent to model fm2. log(chol(((summary(fm1)$sigma)^2)*solve( matrix(getVarCov(fm1), nrow=2 [,1][,2] [1,] -0.3424971 1.492037 [2,] -Inf 1.755388 log(chol(((summary(fm2)$sigma)^2)*solve( matrix(getVarCov(fm2), nrow=2 [,1][,2] [1,] -0.3424971 1.492037 [2,] -Inf 1.755388 In the two mdels, this terms are equals to the optimized parameters in fm2 not in fm1. I am missing something I suppose. Bests, Bernard - [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Multiple stacked barplots on the same graph?
And qplot(x=Categorie,y=Total,data=mydata,geom=bar,fill=Part) + coord_flip() makes it a bit easier to read the labels. Hadley On Dec 5, 2007 8:33 AM, Domenico Vistocco [EMAIL PROTECTED] wrote: This command works: qplot(x=Categorie,y=Total,data=mydata,geom=bar,fill=Part) for your data. domenico vistocco Stéphane CRUVEILLER wrote: Hi, the same error message is displayed with geom=bar as parameter. here is the output of dput: dput(mydata) structure(list(Categorie = structure(c(1L, 12L, 8L, 2L, 5L, 7L, 16L, 6L, 15L, 11L, 10L, 13L, 14L, 3L, 4L, 9L, 17L, 1L, 12L, 8L, 2L, 5L, 7L, 16L, 6L, 15L, 11L, 10L, 13L, 14L, 3L, 4L, 9L, 17L ), .Label = c(Amino acid biosynthesis, Biosynthesis of cofactors, prosthetic groups, and carriers, Cell envelope, Cellular processes, Central intermediary metabolism, DNA metabolism, Energy metabolism, Fatty acid and phospholipid metabolism, Mobile and extrachromosomal element functions, Protein fate, Protein synthesis, Purines, pyrimidines, nucleosides, and nucleotides, Regulatory functions, Signal transduction, Transcription, Transport and binding proteins, Unknown function), class = factor), Part = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c(common, specific), class = factor), Total = c(3.03, 1.65, 1.52, 2.85, 3.4, 11.81, 10.51, 1.95, 2.08, 2.51, 2.23, 7.63, 1.88, 2.76, 7.21, 1.08, 20.75, 0.35, 0.17, 0.08, 0.18, 0.42, 2.05, 1.98, 0.63, 0.17, 0.2, 0.3, 1.58, 0.27, 0.83, 1.38, 3.56, 11.63), chr1 = c(4.55, 2.37, 1.77, 4.68, 3.19, 12.49, 13.56, 2.81, 3.13, 4.58, 3.26, 7.3, 2.06, 3.41, 7.9, 0.22, 22.45, 0.16, 0.06, 0.09, 0.19, 0.09, 0.7, 0.85, 0.22, 0.06, 0.03, 0.32, 0.66, 0.06, 0.63, 0.38, 1.14, 6.17), chr2 = c(1.68, 1.06, 1.55, 1.02, 4.57, 13.87, 7.85, 0.98, 1.06, 0.27, 1.2, 9.88, 2.13, 2.53, 7.71, 0.4, 22.38, 0.71, 0.35, 0.09, 0.22, 0.98, 3.9, 3.24, 0.22, 0.22, 0.49, 0.31, 2.79, 0.62, 1.33, 1.95, 0.44, 16), pl = c(0, 0, 0, 0, 0, 0.17, 4.27, 1.03, 0.34, 0, 0.68, 0.68, 0, 0.17, 1.54, 8.38, 5.3, 0, 0, 0, 0, 0, 2.22, 3.25, 4.44, 0.51, 0, 0.17, 1.88, 0, 0, 4.62, 28.72, 24.27)), .Names = c(Categorie, Part, Total, chr1, chr2, pl), class = data.frame, row.names = c(NA, -34L)) thx, Stéphane. hadley wickham wrote: On Dec 4, 2007 10:34 AM, Stéphane CRUVEILLER [EMAIL PROTECTED] wrote: Hi, I tried this method but it seems that there is something wrong with my data frame: when I type in: qplot(x=as.factor(Categorie),y=Total,data=mydata) It displays a graph with 2 points in each category... but if I add the parameter geom=histogram qplot(x=as.factor(Categorie),y=Total,data=mydata,geom=histogram) Error in storage.mode(test) - logical : object y not found any hint about this... Could you copy and paste the output of dput(mydata) ? (And I'd probably write the plot call as: qplot(Categorie, Total, data=mydata, geom=bar), since it is a bar plot, not a histogram) -- http://had.co.nz/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Plotting error bars in xy-direction
Hans W Borchers wrote: Dear R-help, I am looking for a function that will plot error bars in x- or y-direction (or both), the same as the Gnuplot function 'plot' can achieve with: plot file.dat with xyerrorbars,... Rsite-searching led me to the functions 'errbar' and 'plotCI' in the Hmisc, gregmisc, and plotrix packages. As I understand the descriptions and examples, none of these functions provides horizontal error bars. Looking into 'errbar' and using segments, I wrote a small function for myself adding these kinds of error bars to existing plots. I would still be interested to know what the standard R solution is. Regards, Hans Werner plotCI from plotrix will do horizontal error bars -- from ?plotCI: err: The direction of error bars: x for horizontal, y for vertical (xy would be nice but is not implemented yet; don't know quite how everything would be specified. See examples for composing a plot with simultaneous horizontal and vertical error bars) Ben Bolker -- View this message in context: http://www.nabble.com/Plotting-error-bars-in-xy-direction-tf4948535.html#a14174151 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
8rino-Luca Pantani wrote: Dear R-users. I eventually bought myself a new computer with the following characteristics: Processor AMD ATHLON 64 DUAL CORE 4000+ (socket AM2) Mother board ASR SK-AM2 2 Ram Corsair Value 1 GB DDR2 800 Mhz Hard Disk WESTERN DIGITAL 160 GB SATA2 8MB I'm a newcomer to the Linux world. I started using it (Ubuntu 7.10 at work and FC4 on laptop) on a regular basis on May. I must say I'm quite comfortable with it, even if I have to re-learn a lot of things. But this is not a problem, I will improve my knowledge with time. My main problem now, is that I installed Ubuntu 7.10 Gutsy Gibbon on the new one amd64. To install R on it i followed the directions found here http://help.nceas.ucsb.edu/index.php/Installing_R_on_Ubuntu but unfortunately it did not work. After reading some posts on the R-SIG-debian list, such as https://stat.ethz.ch/pipermail/r-sig-debian/2007-October/000253.html I immediately realize that an amd64 is not the right processor to make life easy. Therefore I would like to know from you, how can I solve this problem: Should I install the i386 version of R ? Should I install another flavour of Linux ? Which one ? Fedora Core 7 ? Debian ? Thanks a lot, for any suggestion Hi, I've got an Athlon 64bits 3000+ processor and Ubuntu LTS (dapper) installed (64 bits version) on my laptop. I do not have any problem to install R from the sources, as long as the correct libraries/compilers/etc. are installed. But it is no pain if you just follow what the configure script tells you (and use apt-get to install missing packages). I guess a common mistake is to forget to install -dev versions of packages, which sometimes contain required headers. However, you should not have troubles installing R on different R distributions, 64bits or not. Hope this help. Thibaut, 64bit-linux-Ruser and still alive. -- ## Thibaut JOMBART CNRS UMR 5558 - Laboratoire de Biométrie et Biologie Evolutive Universite Lyon 1 43 bd du 11 novembre 1918 69622 Villeurbanne Cedex Tél. : 04.72.43.29.35 Fax : 04.72.43.13.88 [EMAIL PROTECTED] http://lbbe.univ-lyon1.fr/-Jombart-Thibaut-.html?lang=en http://pbil.univ-lyon1.fr/software/adegenet/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Which Linux OS on Athlon amd64, to comfortably run R?
Dear R-users. I eventually bought myself a new computer with the following characteristics: Processor AMD ATHLON 64 DUAL CORE 4000+ (socket AM2) Mother board ASR SK-AM2 2 Ram Corsair Value 1 GB DDR2 800 Mhz Hard Disk WESTERN DIGITAL 160 GB SATA2 8MB I'm a newcomer to the Linux world. I started using it (Ubuntu 7.10 at work and FC4 on laptop) on a regular basis on May. I must say I'm quite comfortable with it, even if I have to re-learn a lot of things. But this is not a problem, I will improve my knowledge with time. My main problem now, is that I installed Ubuntu 7.10 Gutsy Gibbon on the new one amd64. To install R on it i followed the directions found here http://help.nceas.ucsb.edu/index.php/Installing_R_on_Ubuntu but unfortunately it did not work. After reading some posts on the R-SIG-debian list, such as https://stat.ethz.ch/pipermail/r-sig-debian/2007-October/000253.html I immediately realize that an amd64 is not the right processor to make life easy. Therefore I would like to know from you, how can I solve this problem: Should I install the i386 version of R ? Should I install another flavour of Linux ? Which one ? Fedora Core 7 ? Debian ? Thanks a lot, for any suggestion -- Ottorino-Luca Pantani, Università di Firenze Dip. Scienza del Suolo e Nutrizione della Pianta P.zle Cascine 28 50144 Firenze Italia Tel 39 055 3288 202 (348 lab) Fax 39 055 333 273 [EMAIL PROTECTED] http://www4.unifi.it/dssnp/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
Note that Ottorino has only 1GB of RAM installed, which makes a 64-bit version of R somewhat moot. See chapter 8 of http://cran.r-project.org/doc/manuals/R-admin.html I would install a i386 version of R on x86_64 Linux unless I had 2Gb or more of RAM. I don't know how easily that works on Ubuntu these days, but I would try it. On Wed, 5 Dec 2007, Ljubomir J. Buturovic wrote: Hi Ottorino, I have been using R on 64-bit Ubuntu for about a year without problems, both Intel and AMD CPUs. Installing and using several packages (e1071, svmpath, survival) also works. However, I had to install R from source: $ gunzip -c R-2.6.1.tar.gz | tar xvf - $ cd R-2.6.1 $ ./configure --enable-R-shlib; make; make pdf # make install; make install-pdf Notice that `make install' has to be run as root. I am using Feisty Fawn (Ubuntu 7.04), although I doubt that makes a difference. Hope this helps, Ljubomir 8rino-Luca Pantani writes: Dear R-users. I eventually bought myself a new computer with the following characteristics: Processor AMD ATHLON 64 DUAL CORE 4000+ (socket AM2) Mother board ASR SK-AM2 2 Ram Corsair Value 1 GB DDR2 800 Mhz Hard Disk WESTERN DIGITAL 160 GB SATA2 8MB I'm a newcomer to the Linux world. I started using it (Ubuntu 7.10 at work and FC4 on laptop) on a regular basis on May. I must say I'm quite comfortable with it, even if I have to re-learn a lot of things. But this is not a problem, I will improve my knowledge with time. My main problem now, is that I installed Ubuntu 7.10 Gutsy Gibbon on the new one amd64. To install R on it i followed the directions found here http://help.nceas.ucsb.edu/index.php/Installing_R_on_Ubuntu but unfortunately it did not work. After reading some posts on the R-SIG-debian list, such as https://stat.ethz.ch/pipermail/r-sig-debian/2007-October/000253.html I immediately realize that an amd64 is not the right processor to make life easy. Therefore I would like to know from you, how can I solve this problem: Should I install the i386 version of R ? Should I install another flavour of Linux ? Which one ? Fedora Core 7 ? Debian ? Thanks a lot, for any suggestion -- Ottorino-Luca Pantani, Università di Firenze Dip. Scienza del Suolo e Nutrizione della Pianta P.zle Cascine 28 50144 Firenze Italia Tel 39 055 3288 202 (348 lab) Fax 39 055 333 273 [EMAIL PROTECTED] http://www4.unifi.it/dssnp/ -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] R function for percentrank
I'm coming late to this, but this *does* need a correction just for the archives ! MS == Marc Schwartz [EMAIL PROTECTED] on Sat, 01 Dec 2007 13:33:21 -0600 writes: MS On Sat, 2007-12-01 at 18:40 +, David Winsemius wrote: David Winsemius [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: tom soyer [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: John, The Excel's percentrank function works like this: if one has a number, x for example, and one wants to know the percentile of this number in a given data set, dataset, one would type =percentrank(dataset,x) in Excel to calculate the percentile. So for example, if the data set is c(1:10), and one wants to know the percentile of 2.5 in the data set, then using the percentrank function one would get 0.166, i.e., 2.5 is in the 16.6th percentile. I am not sure how to program this function in R. I couldn't find it as a built-in function in R either. It seems to be an obvious choice for a built-in function. I am very surprised, but maybe we both missed it. My nomination for a function with a similar result would be ecdf(), the empirical cumulative distribution function. It is of class function so efforts to index ecdf(.)[.] failed for me. I think you did not understand ecdf() !!! It *returns* a function, that you can then apply to old (or new) data; see below MS You can use ls.str() to look into the function environment: ls.str(environment(ecdf(x))) MS f : num 0 MS method : int 2 MS n : int 25 MS x : num [1:25] -2.215 -1.989 -0.836 -0.820 -0.626 ... MS y : num [1:25] 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4 ... MS yleft : num 0 MS yright : num 1 MS You can then use get() or mget() within the function environment to MS return the requisite values. Something along the lines of the following MS within the function percentrank(): MS percentrank - function(x, val) MS { MS env.x - environment(ecdf(x)) MS res - mget(c(x, y), env.x) MS Ind - which(sapply(seq(length(res$x)), MS function(i) isTRUE(all.equal(res$x[i], val MS res$y[Ind] MS } sorry Marc, but Yuck !! - this percentrank() only works when you apply it to original x[i] values - only works for 'val' of length 1 - is a complicated hack and absolutely unneeded (see below) MS Thus: MS set.seed(1) MS x - rnorm(25) x MS [1] -0.62645381 0.18364332 -0.83562861 1.59528080 0.32950777 MS [6] -0.82046838 0.48742905 0.73832471 0.57578135 -0.30538839 MS [11] 1.51178117 0.38984324 -0.62124058 -2.21469989 1.12493092 MS [16] -0.04493361 -0.01619026 0.94383621 0.82122120 0.59390132 MS [21] 0.91897737 0.78213630 0.07456498 -1.98935170 0.61982575 percentrank(x, 0.48742905) MS [1] 0.56 [gives 0.52 in my version of R ] Well, that is *THE SAME* as using ecdf() the way you should have used it : ecdf(x)(0.48742905) {in two lines, that is mypercR - ecdf(x) mypercR(0.48742905) which maybe easier to understand, if you have never used the nice concept that underlies all of approxfun(), splinefun() or ecdf() } You can also use ecdf(x)(x) and indeed check that it is identical to the convoluted percentrank() function above : ecdf(x)(0.48742905) [1] 0.52 ecdf(x)(x) [1] 0.20 0.44 0.12 1.00 0.48 0.16 0.56 0.72 0.60 0.28 0.96 0.52 0.24 0.04 0.92 [16] 0.32 0.36 0.88 0.80 0.64 0.84 0.76 0.40 0.08 0.68 all(ecdf(x)(x) == sapply(x, function(v) percentrank(x,v))) [1] TRUE Regards (and apologies for my apparent indignation ;-) by the author of ecdf() , Martin Maechler, ETH Zurich MS One other approach, which returns the values and their respective rank MS percentiles is: cumsum(prop.table(table(x))) [.. snip ] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Java parser for R data file?
Hi everyone, Has anyone written a parser in Java for either the ASCII or binary format produced by save()? I need to parse a single large 2D array that is structured like this: list( 32609_1 = c(-9549.39231289146, -9574.07159324482, ... ), 32610_2 = c(-6369.12526971635, -6403.99620977124, ... ), 32618_2 = c(-2138.29095689061, -2057.9229403233, ... ), ... ) Or, given that I'm dealing with just a single array, would it be better to roll my own I/O using write.table or write.matrix from the MASS package? Thanks, David __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
Prof Brian Ripley wrote: Note that Ottorino has only 1GB of RAM installed, which makes a 64-bit version of R somewhat moot. See chapter 8 of http://cran.r-project.org/doc/manuals/R-admin.html Only somewhat. The Opteron actually has 1GB too (Hey, it was bought in 2004! And the main point was to see whether 64 bit builds would work at all) but 16GB of swap. So large data sets will fit but be processed slowly. I would install a i386 version of R on x86_64 Linux unless I had 2Gb or more of RAM. I don't know how easily that works on Ubuntu these days, but I would try it. It's not like the 64 bit build feels slow for basic usage, though. I don't think you need to bother with mixing architectures unless you have applications where it really matters (CPU intensive, but below 32-bit addressing limitations). Buying more RAM is much to be preferred. Now what kind of RAM does my Opteron board take... ? -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
On Wed, Dec 05, 2007 at 06:11:40PM +0100, Peter Dalgaard wrote: One oddity about Ubuntu is that there are no CRAN builds for 64bit. Volunteers would be welcomed with open arms. Dirk, -- Three out of two people have difficulties with fractions. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] confidence intervals for y predicted in non linear regression
Prof Brian Ripley wrote: You mean the package nls2 at http://w3.jouy.inra.fr/unites/miaj/public/AB/nls2/install.html and not the unfortunately named nls2 that has just appeared on CRAN? The first is not really a 'package' in the R sense. Actually, it is one (sort of), but it is broken. The instructions say that you can use R CMD INSTALL to install in R 2.0.0 (!) With a current R, you can try but it dies: No man pages found in package 'nls2' ** building package indices ... Warning in file(file, r, encoding = encoding) : cannot open file '../R/init.R', reason 'No such file or directory' Error in file(file, r, encoding = encoding) : unable to open connection Calls: Anonymous ... switch - sys.source - eval - eval - source - file Execution halted ERROR: installing package indices failed ** Removing '/home/bs/pd/Rlibrary/nls2' ...and there were some odd goings-on at the start as well. -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Working with ts objects
I am relatively new to R and object oriented programming. I have relied on SAS for most of my data analysis. I teach an introductory undergraduate forecasting course using the Diebold text and I am considering using R in addition to SAS and Eviews in the course. I work primarily with univariate or multivariate time series data. I am having a great deal of difficulty understanding and working with ts objects particularly when it comes to referencing variables in plot commands or in formulas. The confusion is amplified when certain procedures (lm for example) coerce the ts object into a data.frame before application with the results that the output is stored in a data.frame object. For example the two sets of code below replicate examples from chapter 2 and 6 in the text. In the first set of code if I were to replace anscombe-read.table(fname, header=TRUE) with anscombe-ts(read.table(fname, header=TRUE)) the plot() commands would generate errors. The objects x1, y1 ... would not be recognized. In this case I would have to reference the specific column in the anscombe data set. If I would have constructed the data set from several different data sets using the ts.intersect() function (see second code below)the problem becomes even more involved and keeping track of which columns are associated with which variables can be rather daunting. All I wanted was to plot actual vs. predicted values of hstarts and the residuals from the model. Given the difficulties I have encountered I know my students will have similar problems. Is there a source other than the basic R manuals that I can consult and recommend to my students that will help get a handle on working with time series objects? I found the Shumway Time series analysis and its applications with R Examples website very helpful but many practical questions involving manipulation of time series data still remain. Any help will be appreciated. Thanks, Richard Saba Department of Economics Auburn University Email: [EMAIL PROTECTED] Phone: 334 844-2922 anscombe-read.table(fname, header=TRUE) names(anscombe)-c(x1,y1,x2,y2,x3,y3,x4,y4) reg1-lm(y1~1 + x1, data=anscombe) reg2-lm(y2~1 + x2, data=anscombe) reg3-lm(y3~1 + x3, data=anscombe) reg4-lm(y4~1 + x4, data=anscombe) summary(reg1) summary(reg2) summary(reg3) summary(reg4) par(mfrow=c(2,2)) plot(x1,y1) abline(reg1) plot(x2,y2) abline(reg2) plot(x3,y3) abline(reg3) plot(x4,y4) abline(reg4) .. fname-file.choose() tab6.1-ts(read.table(fname, header=TRUE),frequency=12,start=c(1946,1)) month-cycle(tab6.1) year-floor(time(tab6.1)) dat1-ts.intersect(year,month,tab6.1) dat2-window(dat1,start=c(1946,1),end=c(1993,12)) reg1-lm(tab6.1~1+factor(month),data=dat2, na.action=NULL) summary(reg1) hstarts-dat2[,3] plot1-ts.intersect(hstarts,reg1$fitted.value,reg1$resid) plot.ts(plot1[,1]) lines(plot1[,2], col=red) plot.ts(plot[,3], ylab=Residuals) __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
On Wed, 05-Dec-2007 at 06:11PM +0100, Peter Dalgaard wrote: [] | One oddity about Ubuntu is that there are no CRAN builds for 64bit. | Presumably, the Debian packages work, or you can get the build | script and make your own build. This is not really within my | domain, though. I've always used rpms (or debs for Debian-type OSes) but I install R from source which is very easy to do. Adding R packages with install.packages() is also extremely easy. If you have the 64bit OS, it will compile R as 64 bit (enless you make some modifications to the standard configuration). One downside of that is that you'd be unable to use packages that have only 32 bit versions. One such is ASReml-R but if you never intend to use such things, the only other consideration I can think of is the relatively small amount of memory. No great benefits of 64 bit without lots of memory, but a few downsides. HTH -- ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~. ___Patrick Connolly {~._.~} Great minds discuss ideas _( Y )_Middle minds discuss events (:_~*~_:)Small minds discuss people (_)-(_) . Anon ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Java parser for R data file?
On Wed, 5 Dec 2007, David Coppit wrote: Hi everyone, Has anyone written a parser in Java for either the ASCII or binary format produced by save()? I need to parse a single large 2D array that is structured like this: list( 32609_1 = c(-9549.39231289146, -9574.07159324482, ... ), 32610_2 = c(-6369.12526971635, -6403.99620977124, ... ), 32618_2 = c(-2138.29095689061, -2057.9229403233, ... ), ... ) Or, given that I'm dealing with just a single array, would it be better to roll my own I/O using write.table or write.matrix from the MASS package? It would be much easier. The save() format is far more complex than you need. However, I would use writeBin() to write a binary file and read that in in Java, avoiding the binary - ASCII - binary conversion. -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
8rino-Luca Pantani wrote: Dear R-users. I eventually bought myself a new computer with the following characteristics: Processor AMD ATHLON 64 DUAL CORE 4000+ (socket AM2) Mother board ASR SK-AM2 2 Ram Corsair Value 1 GB DDR2 800 Mhz Hard Disk WESTERN DIGITAL 160 GB SATA2 8MB I'm a newcomer to the Linux world. I started using it (Ubuntu 7.10 at work and FC4 on laptop) on a regular basis on May. I must say I'm quite comfortable with it, even if I have to re-learn a lot of things. But this is not a problem, I will improve my knowledge with time. My main problem now, is that I installed Ubuntu 7.10 Gutsy Gibbon on the new one amd64. To install R on it i followed the directions found here http://help.nceas.ucsb.edu/index.php/Installing_R_on_Ubuntu but unfortunately it did not work. After reading some posts on the R-SIG-debian list, such as https://stat.ethz.ch/pipermail/r-sig-debian/2007-October/000253.html I immediately realize that an amd64 is not the right processor to make life easy. Therefore I would like to know from you, how can I solve this problem: Should I install the i386 version of R ? Should I install another flavour of Linux ? Which one ? Fedora Core 7 ? Debian ? Thanks a lot, for any suggestion Amd64 architecture should not be a major issue for R on any of the major platforms, as far as I know. I have Fedora 7 (soon-ish F8) on the big machine back home (dual Opteron) and that one never had any major issues with either of source builds or the official RPMs. The main (only?) thing that still tends to bite people on 64 bit is browser plugins, notably Java. In general, I've been happy with Fedora, although its desire to update itself constantly does require a good 'Net connection. My SUSE desktop at work is also 64 bit and happy to deal with R (in fact the official release builds are made on it). The KDE desktop has a few annoying misfeatures (to me), though, and you need a little special setup to include Detlefs repository as an install source. One oddity about Ubuntu is that there are no CRAN builds for 64bit. Presumably, the Debian packages work, or you can get the build script and make your own build. This is not really within my domain, though. -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] newbie lapply question
On Wed, 5 Dec 2007, Prof Brian Ripley wrote: [...] Thanks I'll read it more carefully. Perhaps if you told us what you are trying to achieve we might be able to help you achieve it. I have a function which takes a date as an argument. I've tested it, and I'd like to run it over a range of dates. So I'm looking at apply- or map- type functions. -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Dealing with NA's in a data matrix
Hi I have a matrix with NA value that I would like to convert these to a value of 0. any suggestions Kind Regards Amit Patel ___ ttp://uk.promotions.yahoo.com/forgood/ [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Displaying numerics to full double precision
Jeff Delmerico wrote: I'm working on a shared library of C functions for use with R, and I want to create a matrix in R and pass it to the C routines. I know R computes and supposedly stores numerics in double precision, but when I create a matrix of random numerics using rnorm(), the values are displayed in single precision, and also exported in single precision when I pass them out to my C routines. An example is below: a - matrix(rnorm(16, mean=10, sd=4), nrow=4) a [,1] [,2] [,3] [,4] [1,] 14.907606 17.572872 19.708977 9.809943 [2,] 9.322041 13.624452 7.745254 7.596176 [3,] 10.642408 6.151546 9.937434 6.913875 [4,] 14.617647 5.577073 8.217559 12.115465 storage.mode(a) [1] double Does anyone know if there is a way to change the display or storage settings so that the values will be displayed to their full precision? Or does rnorm only produce values to single precision? Any assistance would be greatly appreciated. Thanks, Jeff Delmerico options(digits) # 7 options(digits=x) I may be mistaken, but I think the values are indeed exported as double precision -- the issue here is just a display setting. Ben Bolker -- View this message in context: http://www.nabble.com/Displaying-numerics-to-full-double-precision-tf4950807.html#a14178707 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Is R portable?
Peter Dalgaard wrote: The toolchain availability tends to get in the way. Linux-based gadgets could prove easier. I do wonder from time to time whether there really is a market for R on cellphones... As soon as someone writes library(ringtone) there might be :) And I think you'd have to turn off predictive text. Can someone with a mobile/cellphone tell me what 'hist(runif(100))' comes up as? [1] Barry [1] No, I haven't got one. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Is R portable?
On 05-Dec-07 18:57:58, Barry Rowlingson wrote: Peter Dalgaard wrote: The toolchain availability tends to get in the way. Linux-based gadgets could prove easier. I do wonder from time to time whether there really is a market for R on cellphones... As soon as someone writes library(ringtone) there might be :) And I think you'd have to turn off predictive text. Can someone with a mobile/cellphone tell me what 'hist(runif(100))' comes up as? [1] Barry [1] No, I haven't got one. I have a very old cellphone whose display, when I switch it on, looks very much like what one would expect from 'hist(runif(100))'. I'm not using it any more. Ted. E-Mail: (Ted Harding) [EMAIL PROTECTED] Fax-to-email: +44 (0)870 094 0861 Date: 05-Dec-07 Time: 19:32:42 -- XFMail -- __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] 2/3d interpolation from a regular grid to another regular grid
I just read the description in ?Krig in the package fields which says: Fits a surface to irregularly spaced data. Yes, that is the most general case. Regular data location is a subset of irregular. Anyway, kriging, just one g, after the name of Danie Krige, the south african statistician who first applied such method for minig survey. My problem is simpler ... So it is really purely numerical. ... I just hoped that R had that already coded ... Of course R has ... ;) If your grids are really as simple as the example you posted above, and you have a really little variability, all you need is a moving average, the arithmetic mean of the two nearest points belonging to grid1 and grid2 respectively. I assume that your regularly shaped grids are values stored in matrix objects. The functions comes from the diff.default code (downloading the R source code, I assure, is worth): my.interp - function(x, lag = 1) { r - unclass(x) # don't want class-specific subset methods i1 - -1:-lag r - (r[i1] + r[-length(r):-(length(r)-lag+1)])/2 class(r) - oldClass(x) return(r) } Finally, g1 - apply(grid1val,1,my.interp) g2 - apply(grid2val,2,my.interp) give the interpolations on gridFinal, provided that all gridFinal points are within the grid1 and grid2 ones. If you want the mean from 4 points, you apply once more with lag=3, cbind/rbind to the result columns/rows o NAs, and you calculate the mean of the points of the two matrixes. This is the simplest (and quickest) moving average that you can do. For more complicated examples, and for 3d, you have to go a little further, but the principle holds. ScionForbai __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Interpretation of 'Intercept' in a 2-way factorial lm
Hi all, I hope this question is not too trivial. I can't find an explanation anywhere (Stats and R books, R-archives) so now I have to turn to the R-list. Question: If you have a factorial design with two factors (say A and B with two levels each). What does the intercept coefficient with treatment.contrasts represent?? Here is an example without interaction where A has two levels A1 and A2, and B has two levels B1 and B2. So R takes as a baseline A1 and B1. coef( summary ( lm ( fruit ~ A + B, data = test))) Estimate Std. Error t value Pr(|t|) (Intercept) 2.716667 0.5484828 4.953058 7.879890e-04 A26.27 0.633 9.894737 3.907437e-06 B25.17 0.633 8.157895 1.892846e-05 I understand that the mean of A2 is +6.3 more than A1, and that B2 is 5.2 more than B1. So the question is: Is the intercept A1 and B1 combined as one mean (the baseline)? or is it something else? Does this number actually tell me anything useful (2.716)?? What does the model (y = intercept + ??) look like then? I can't understand how both factors (A and B) can have the same intercept? Thanks in advance!! Gustaf Granath Dept of Plant Ecology Uppsala University, Sweden __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] newbie lapply question
I am not sure to understand your problem, but it seems to me that you can use directly the function on the range of the dates: x=as.Date(c('2007-01-01','2007-01-02')) fff=function(x){y=x+1;return(y)} fff(x) [1] 2007-01-02 2007-01-03 class(fff(x)) [1] Date Perhaps your function use a different input (not a vector of dates but a dataframe)? domenico vistocco Ranjan Bagchi wrote: On Wed, 5 Dec 2007, Prof Brian Ripley wrote: [...] Thanks I'll read it more carefully. Perhaps if you told us what you are trying to achieve we might be able to help you achieve it. I have a function which takes a date as an argument. I've tested it, and I'd like to run it over a range of dates. So I'm looking at apply- or map- type functions. -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] how to interpolate a plot with a logistic curve
hello, I have this simple question. This is my dataset size 1 57 2 97 3 105 4 123 5 136 6 153 7 173 8 180 9 193 10 202 11 213 12 219 13 224 14 224 15 248 16 367 17 496 18 568 19 618 20 670 21 719 22 774 23 810 24 814 25 823 I plot it with: plot(generalstats[,1], type=b, xlab=Mesi, ylab=Numero di vertici, main=); and try to interpolate with a linear regression with abline(lm(generalstats[, 1 ]~ c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25)), lty=3, col=red); how to interpolate the data with a logistic curve? I cannot find the - I suppose easy - solution.. thank you, Simone __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] how to interpolate a plot with a logistic curve
On Wednesday 05 December 2007, Simone Gabbriellini wrote: hello, I have this simple question. This is my dataset size 1 57 2 97 3 105 4 123 5 136 6 153 7 173 8 180 9 193 10202 11213 12219 13224 14224 15248 16367 17496 18568 19618 20670 21719 22774 23810 24814 25823 I plot it with: plot(generalstats[,1], type=b, xlab=Mesi, ylab=Numero di vertici, main=); and try to interpolate with a linear regression with abline(lm(generalstats[, 1 ]~ c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25)), lty=3, col=red); how to interpolate the data with a logistic curve? I cannot find the - I suppose easy - solution.. thank you, Simone try: glm(formula, data, family=binomial()) require(Design) lrm() Cheers, -- Dylan Beaudette Soil Resource Laboratory http://casoilresource.lawr.ucdavis.edu/ University of California at Davis 530.754.7341 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] R function for percentrank
On Wed, 2007-12-05 at 18:42 +0100, Martin Maechler wrote: I'm coming late to this, but this *does* need a correction just for the archives ! MS == Marc Schwartz [EMAIL PROTECTED] on Sat, 01 Dec 2007 13:33:21 -0600 writes: MS On Sat, 2007-12-01 at 18:40 +, David Winsemius wrote: David Winsemius [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: tom soyer [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: John, The Excel's percentrank function works like this: if one has a number, x for example, and one wants to know the percentile of this number in a given data set, dataset, one would type =percentrank(dataset,x) in Excel to calculate the percentile. So for example, if the data set is c(1:10), and one wants to know the percentile of 2.5 in the data set, then using the percentrank function one would get 0.166, i.e., 2.5 is in the 16.6th percentile. I am not sure how to program this function in R. I couldn't find it as a built-in function in R either. It seems to be an obvious choice for a built-in function. I am very surprised, but maybe we both missed it. My nomination for a function with a similar result would be ecdf(), the empirical cumulative distribution function. It is of class function so efforts to index ecdf(.)[.] failed for me. I think you did not understand ecdf() !!! It *returns* a function, that you can then apply to old (or new) data; see below MS You can use ls.str() to look into the function environment: ls.str(environment(ecdf(x))) MS f : num 0 MS method : int 2 MS n : int 25 MS x : num [1:25] -2.215 -1.989 -0.836 -0.820 -0.626 ... MS y : num [1:25] 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 0.36 0.4 ... MS yleft : num 0 MS yright : num 1 MS You can then use get() or mget() within the function environment to MS return the requisite values. Something along the lines of the following MS within the function percentrank(): MS percentrank - function(x, val) MS { MS env.x - environment(ecdf(x)) MS res - mget(c(x, y), env.x) MS Ind - which(sapply(seq(length(res$x)), MS function(i) isTRUE(all.equal(res$x[i], val MS res$y[Ind] MS } sorry Marc, but Yuck !! - this percentrank() only works when you apply it to original x[i] values - only works for 'val' of length 1 - is a complicated hack and absolutely unneeded (see below) MS Thus: MS set.seed(1) MS x - rnorm(25) x MS [1] -0.62645381 0.18364332 -0.83562861 1.59528080 0.32950777 MS [6] -0.82046838 0.48742905 0.73832471 0.57578135 -0.30538839 MS [11] 1.51178117 0.38984324 -0.62124058 -2.21469989 1.12493092 MS [16] -0.04493361 -0.01619026 0.94383621 0.82122120 0.59390132 MS [21] 0.91897737 0.78213630 0.07456498 -1.98935170 0.61982575 percentrank(x, 0.48742905) MS [1] 0.56 [gives 0.52 in my version of R ] Well, that is *THE SAME* as using ecdf() the way you should have used it : ecdf(x)(0.48742905) {in two lines, that is mypercR - ecdf(x) mypercR(0.48742905) which maybe easier to understand, if you have never used the nice concept that underlies all of approxfun(), splinefun() or ecdf() } You can also use ecdf(x)(x) and indeed check that it is identical to the convoluted percentrank() function above : ecdf(x)(0.48742905) [1] 0.52 ecdf(x)(x) [1] 0.20 0.44 0.12 1.00 0.48 0.16 0.56 0.72 0.60 0.28 0.96 0.52 0.24 0.04 0.92 [16] 0.32 0.36 0.88 0.80 0.64 0.84 0.76 0.40 0.08 0.68 all(ecdf(x)(x) == sapply(x, function(v) percentrank(x,v))) [1] TRUE Regards (and apologies for my apparent indignation ;-) by the author of ecdf() , Martin Maechler, ETH Zurich Martin, Thanks for the corrections. In hindsight, now seeing the intended use of ecdf() in the fashion you describe above, it is now clear that my approach in response to David's query was un-needed and over the top. Yuck is quite appropriate... :-) As I was going through this exercise, it did seem overly complicated, given R's usual elegant philosophy about such things. I suppose if I had looked at the source for plot.stepfun(), it would have been more evident as to how the y values are acquired. In reviewing the examples in ?ecdf, I think that an example using something along the lines of the discussion here more explicitly, would be helpful. It is not crystal clear from the examples, that one can use ecdf() in this fashion, though the use of 12 * Fn(tt) hints at it. Perhaps: ##-- Simple didactical ecdf example: x - rnorm(12) Fn - ecdf(x) Fn Fn(x) # returns the percentiles for x ... Thanks again Martin and no offense taken... :-) Regards, Marc
Re: [R] kalman filter random walk
You may want to look at package dlm. Giovanni Date: Wed, 05 Dec 2007 12:05:00 -0600 From: Alexander Moreno [EMAIL PROTECTED] Sender: [EMAIL PROTECTED] Precedence: list Hi, I'm trying to use the kalman filter to estimate the variable drift of a random walk, given that I have a vector of time series data. Anyone have any thoughts on how to do this in R? Thanks, Alex [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Giovanni Petris [EMAIL PROTECTED] Department of Mathematical Sciences University of Arkansas - Fayetteville, AR 72701 Ph: (479) 575-6324, 575-8630 (fax) http://definetti.uark.edu/~gpetris/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Which Linux OS on Athlon amd64, to comfortably run R?
Hi Ottorino, I have been using R on 64-bit Ubuntu for about a year without problems, both Intel and AMD CPUs. Installing and using several packages (e1071, svmpath, survival) also works. However, I had to install R from source: $ gunzip -c R-2.6.1.tar.gz | tar xvf - $ cd R-2.6.1 $ ./configure --enable-R-shlib; make; make pdf # make install; make install-pdf Notice that `make install' has to be run as root. I am using Feisty Fawn (Ubuntu 7.04), although I doubt that makes a difference. Hope this helps, Ljubomir 8rino-Luca Pantani writes: Dear R-users. I eventually bought myself a new computer with the following characteristics: Processor AMD ATHLON 64 DUAL CORE 4000+ (socket AM2) Mother board ASR SK-AM2 2 Ram Corsair Value 1 GB DDR2 800 Mhz Hard Disk WESTERN DIGITAL 160 GB SATA2 8MB I'm a newcomer to the Linux world. I started using it (Ubuntu 7.10 at work and FC4 on laptop) on a regular basis on May. I must say I'm quite comfortable with it, even if I have to re-learn a lot of things. But this is not a problem, I will improve my knowledge with time. My main problem now, is that I installed Ubuntu 7.10 Gutsy Gibbon on the new one amd64. To install R on it i followed the directions found here http://help.nceas.ucsb.edu/index.php/Installing_R_on_Ubuntu but unfortunately it did not work. After reading some posts on the R-SIG-debian list, such as https://stat.ethz.ch/pipermail/r-sig-debian/2007-October/000253.html I immediately realize that an amd64 is not the right processor to make life easy. Therefore I would like to know from you, how can I solve this problem: Should I install the i386 version of R ? Should I install another flavour of Linux ? Which one ? Fedora Core 7 ? Debian ? Thanks a lot, for any suggestion -- Ottorino-Luca Pantani, Universit? di Firenze Dip. Scienza del Suolo e Nutrizione della Pianta P.zle Cascine 28 50144 Firenze Italia Tel 39 055 3288 202 (348 lab) Fax 39 055 333 273 [EMAIL PROTECTED] http://www4.unifi.it/dssnp/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Displaying numerics to full double precision
Jeff Delmerico wrote: Thanks Ben, that fixed the display within R. However, even after changing the display settings, the matrix elements still appear to be exported in single precision. The matrix object is being passed into my C routines as an SEXP Numeric type, and somewhere along the way, some of the digits are getting lost. Here's the relevant bit of my C code: SEXP divideMatrix(SEXP matrix_in, SEXP sub_height, SEXP sub_width, SEXP fileS) ... if ( isMatrix(matrix_in) isNumeric(matrix_in) ) { /* Use R macros to convert from SEXP to C types */ matrix = REAL(matrix_in); height = INTEGER(GET_DIM(matrix_in))[0]; width = INTEGER(GET_DIM(matrix_in))[1]; subW = INTEGER_VALUE(sub_width); subH = INTEGER_VALUE(sub_height); ... } Am I using the wrong macro to convert into a double in C? Any ideas? Thanks, Jeff Delmerico Ben Bolker wrote: Jeff Delmerico wrote: I'm working on a shared library of C functions for use with R, and I want to create a matrix in R and pass it to the C routines. I know R computes and supposedly stores numerics in double precision, but when I create a matrix of random numerics using rnorm(), the values are displayed in single precision, and also exported in single precision when I pass them out to my C routines. An example is below: a - matrix(rnorm(16, mean=10, sd=4), nrow=4) a [,1] [,2] [,3] [,4] [1,] 14.907606 17.572872 19.708977 9.809943 [2,] 9.322041 13.624452 7.745254 7.596176 [3,] 10.642408 6.151546 9.937434 6.913875 [4,] 14.617647 5.577073 8.217559 12.115465 storage.mode(a) [1] double Does anyone know if there is a way to change the display or storage settings so that the values will be displayed to their full precision? Or does rnorm only produce values to single precision? Any assistance would be greatly appreciated. Thanks, Jeff Delmerico options(digits) # 7 options(digits=x) I may be mistaken, but I think the values are indeed exported as double precision -- the issue here is just a display setting. Ben Bolker I'm not sure. I do know that Rinternals.h has #define REAL(x) ((double *) DATAPTR(x)) so that doesn't seem to be the problem ... Ben -- View this message in context: http://www.nabble.com/Displaying-numerics-to-full-double-precision-tf4950807.html#a14179985 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Dealing with NA's in a data matrix
x[is.na(x)] - 0 On 05/12/2007, Amit Patel [EMAIL PROTECTED] wrote: Hi I have a matrix with NA value that I would like to convert these to a value of 0. any suggestions Kind Regards Amit Patel ___ ttp://uk.promotions.yahoo.com/forgood/ [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Henrique Dallazuanna Curitiba-Paraná-Brasil 25° 25' 40 S 49° 16' 22 O __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Interpretation of 'Intercept' in a 2-way factorial lm
Gustaf Granath wrote: Hi all, I hope this question is not too trivial. I can't find an explanation anywhere (Stats and R books, R-archives) so now I have to turn to the R-list. Question: If you have a factorial design with two factors (say A and B with two levels each). What does the intercept coefficient with treatment.contrasts represent?? Here is an example without interaction where A has two levels A1 and A2, and B has two levels B1 and B2. So R takes as a baseline A1 and B1. coef( summary ( lm ( fruit ~ A + B, data = test))) Estimate Std. Error t value Pr(|t|) (Intercept) 2.716667 0.5484828 4.953058 7.879890e-04 A26.27 0.633 9.894737 3.907437e-06 B25.17 0.633 8.157895 1.892846e-05 I understand that the mean of A2 is +6.3 more than A1, and that B2 is 5.2 more than B1. So the question is: Is the intercept A1 and B1 combined as one mean (the baseline)? or is it something else? Does this number actually tell me anything useful (2.716)?? What does the model (y = intercept + ??) look like then? I can't understand how both factors (A and B) can have the same intercept? Consider an AxB crosstable of (fitted) means. Upper left corner is intercept , add A2, B2, or both to get the other three cells. -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] newbie lapply question
Hi -- I just noticed the following (R 2.6.1 on OSX) lapply(c(as.Date('2007-01-01')), I) [[1]] [1] 13514 This is a bit surprising.. Why does lapply unclass the object? Sorry for such a basic question, I don't seem able to produce the right google keywords. Ranjan __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Interpretation of 'Intercept' in a 2-way factorial lm
You estimate a model with the Factors A or B either present (1) or not present (0) and with an intercept. Thus you would predict: For both A and B not present: Intercept For A only present: Intercept+coef(A) For B only preseent: Intercept+coef(B) For both present: Intercept+coef(A)+coef(B). Again, you would interpret the intercept as the value of fruit when A and B are not present (or inactive). If the intercept is not meaningful in your setting and you just want to know if both groups differ, then you want to use function aov I guess. What is your fruit variable? I would also suggest to visually inspect your data. That always helps :) The code is also down below. Look at the following example in which 4 x 10 Ys are drawn randomly from normal distributions with equal variance but different means. The first ten observations have both A and B not present (i.e. 0) as specified in the vectors a and b. The mean of these observations where A and B are zero is 1 as specified in y1=rnorm(10, - 1 -,1). As you will see if you run this code, the estimated Intercept is 1.0512 which is close to 1 (the true mean). As you see (just confirming what was said above), this is the average of the baseline (or reference group if you will) when both A and B are absent. y1=rnorm(10,1,1) y2=rnorm(10,2,1) y3=rnorm(10,3,1) y4=rnorm(10,4,1) a=c(rep(0,20),rep(1,20)) b=c(rep(0,10),rep(1,10),rep(0,10),rep(1,10)) y=c(y1,y2,y3,y4) data=data.frame(cbind(y,a,b)) Plot interaction.plot(a,b,y) Models summary(lm(y~factor(a)+factor(b),data=data) Compare this to summary(aov(y~factor(a)+factor(b),data=data) Cheers, Daniel - cuncta stricte discussurus - -Ursprüngliche Nachricht- Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Im Auftrag von Gustaf Granath Gesendet: Wednesday, December 05, 2007 2:32 PM An: r-help@r-project.org Betreff: [R] Interpretation of 'Intercept' in a 2-way factorial lm Hi all, I hope this question is not too trivial. I can't find an explanation anywhere (Stats and R books, R-archives) so now I have to turn to the R-list. Question: If you have a factorial design with two factors (say A and B with two levels each). What does the intercept coefficient with treatment.contrasts represent?? Here is an example without interaction where A has two levels A1 and A2, and B has two levels B1 and B2. So R takes as a baseline A1 and B1. coef( summary ( lm ( fruit ~ A + B, data = test))) Estimate Std. Error t value Pr(|t|) (Intercept) 2.716667 0.5484828 4.953058 7.879890e-04 A26.27 0.633 9.894737 3.907437e-06 B25.17 0.633 8.157895 1.892846e-05 I understand that the mean of A2 is +6.3 more than A1, and that B2 is 5.2 more than B1. So the question is: Is the intercept A1 and B1 combined as one mean (the baseline)? or is it something else? Does this number actually tell me anything useful (2.716)?? What does the model (y = intercept + ??) look like then? I can't understand how both factors (A and B) can have the same intercept? Thanks in advance!! Gustaf Granath Dept of Plant Ecology Uppsala University, Sweden __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] File based configuration
I'm wanting to run R scripts non-interactively as part of a technology independent framework. I want control over the behaviour of these processes by specifying various global variables in a configuration file that would be passed as a command line argument. I'm wondering if you know of any R support for configuration file formats. (i.e. any functions that would read a configuration file of some common format) For example: -The .properties configuration format for java seems to be quite popular, would I have to read it in by writing some kind of java extension to R? -An XML configuration format could also be possible, but it's overkill for my needs. Any help would be greatly appreciated __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Java parser for R data file?
David Coppit wrote: Hi everyone, Has anyone written a parser in Java for either the ASCII or binary format produced by save()? You might want to consider using the hdf5 package to save the array in HDF5 format. There are HDF5 libraries for Java as well http://hdf.ncsa.uiuc.edu/hdf-java-html/. I have never used them, but it works quite well for transferring data between R and Python. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] coxme frailty model standard errors?
Hello, I am running R 2.6.1 on windows xp I am trying to fit a cox proportional hazard model with a shared Gaussian frailty term using coxme My model is specified as: nofit1-coxme(Surv(Age,cen1new)~ Sex+bo2+bo3,random=~1|isl,data=mydat) With x1-x3 being dummy variables, and isl being the community level variable with 4 levels. Does anyone know if there is a way to get the standard error for the random effect, like in nofit1$var? I would like to know if my random effect is worth writing home about. Any help would be most appreciated Corey Sparks I can get the following output nofit1-coxme(Surv(Age,cen1new)~ Sex+bo2+bo3,random=~1|isl, data=no1901) nofit1 Cox mixed-effects model fit by maximum likelihood Data: no1901 n=959 (2313 observations deleted due to missingness) Iterations= 3 69 NULL Integrated Penalized Log-likelihood -600.0795 -581.1718 -577.9682 Penalized loglik: chisq= 44.22 on 5.61 degrees of freedom, p= 4.3e-08 Integrated loglik: chisq= 37.82 on 4 degrees of freedom, p= 1.2e-07 Fixed effects: Surv(Age, cen1new) ~ Sex + bo2 + bo3 coef exp(coef) se(coef)z p Sex 0.2269214 1.254731 0.2151837 1.05 0.2900 bo2 0.5046991 1.656487 0.2510523 2.01 0.0440 bo3 1.0606144 2.888145 0.2726000 3.89 0.0001 Random effects: ~1 | isl isl Variance: 0.3876189 Corey Sparks Assistant Professor Department of Demography and Organization Studies University of Texas-San Antonio One UTSA Circle San Antonio TX 78249 Phone: 210 458 6858 [EMAIL PROTECTED] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] coxme frailty model standard errors?
Hello, I am running R 2.6.1 on windows xp I am trying to fit a cox proportional hazard model with a shared Gaussian frailty term using coxme My model is specified as: nofit1-coxme(Surv(Age,cen1new)~ Sex+bo2+bo3,random=~1|isl,data=mydat) With x1-x3 being dummy variables, and isl being the community level variable with 4 levels. Does anyone know if there is a way to get the standard error for the random effect, like in nofit1$var? I would like to know if my random effect is worth writing home about. Any help would be most appreciated Corey Sparks I can get the following output nofit1-coxme(Surv(Age,cen1new)~ Sex+bo2+bo3,random=~1|isl, data=no1901) nofit1 Cox mixed-effects model fit by maximum likelihood Data: no1901 n=959 (2313 observations deleted due to missingness) Iterations= 3 69 NULL Integrated Penalized Log-likelihood -600.0795 -581.1718 -577.9682 Penalized loglik: chisq= 44.22 on 5.61 degrees of freedom, p= 4.3e-08 Integrated loglik: chisq= 37.82 on 4 degrees of freedom, p= 1.2e-07 Fixed effects: Surv(Age, cen1new) ~ Sex + bo2 + bo3 coef exp(coef) se(coef)z p Sex 0.2269214 1.254731 0.2151837 1.05 0.2900 bo2 0.5046991 1.656487 0.2510523 2.01 0.0440 bo3 1.0606144 2.888145 0.2726000 3.89 0.0001 Random effects: ~1 | isl isl Variance: 0.3876189 Corey Sparks Assistant Professor Department of Demography and Organization Studies University of Texas-San Antonio One UTSA Circle San Antonio TX 78249 Phone: 210 458 6858 [EMAIL PROTECTED] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] 2/3d interpolation from a regular grid to another regular grid
On 2007-December-05 , at 16:47 , Scionforbai wrote: I just read the description in ?Krig in the package fields which says: Fits a surface to irregularly spaced data. Yes, that is the most general case. Regular data location is a subset of irregular. Anyway, kriging, just one g, after the name of Danie Krige, the south african statistician who first applied such method for minig survey. ooops. sorry about the typo. My problem is simpler ... So it is really purely numerical. ... I just hoped that R had that already coded ... Of course R has ... ;) If your grids are really as simple as the example you posted above, and you have a really little variability, all you need is a moving average, the arithmetic mean of the two nearest points belonging to grid1 and grid2 respectively. I assume that your regularly shaped grids are values stored in matrix objects. The functions comes from the diff.default code (downloading the R source code, I assure, is worth): I can imagine it is indeed. I use the source of packages functions very often. my.interp - function(x, lag = 1) { r - unclass(x) # don't want class-specific subset methods i1 - -1:-lag r - (r[i1] + r[-length(r):-(length(r)-lag+1)])/2 class(r) - oldClass(x) return(r) } Finally, g1 - apply(grid1val,1,my.interp) g2 - apply(grid2val,2,my.interp) give the interpolations on gridFinal, provided that all gridFinal points are within the grid1 and grid2 ones. If you want the mean from 4 points, you apply once more with lag=3, cbind/rbind to the result columns/rows o NAs, and you calculate the mean of the points of the two matrixes. This is the simplest (and quickest) moving average that you can do. For more complicated examples, and for 3d, you have to go a little further, but the principle holds. Thanks very much. I'll test this soon (and it looks like the vector operation might even be directly translatable in Fortran which is nice since I'll need to do it in Fortran too). Thanks again. JiHO --- http://jo.irisson.free.fr/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Learning to do randomized block design analysis
Dear R-Helpers (1) After a night's sleep, I realized why the other helpers think differently from me. I agree with others that it may be better to use multi-stratum model but I was a bit surprised since they seem to think 'block' variable should *not* be a fixed effect. A. Others seemed to think, Kevin is trying to estimate 'multi-stratum' model since he is using Error(block) B. I thought Kevin is trying to estimate a simple ANOVA model (*not* random effects model) but did not use the right R code I thought so for the following reasons. 1) I looked up the book in Amazon and browsed the index using ' Search Inside' function. It does not seem to cover random effects model. 2) The description of the book says it uses Minitab so I guessed Kevin is getting R code from somewhere else. 3) In addition, Kevin's code looked very similar to the example code of 'aov'. The example code of 'aov' has the following code segment involving 'block' variable ## as a test, not particularly sensible statistically npk.aovE - aov(yield ~ N*P*K + Error(block), npk) So, I guessed Kevin might have gotten his code from here. After emailing Kevin, I found that he is using the code from 'split-plot' section of MASS, so my guess is not that far off. (2) I got this new bits of information from Kevin. The data set is from a psychological experiment and subjects are *assigned* to one of the blocks according to their scores on a test. Subjects with the lowest scores are assigned to block A, and highest to block E. These blocks were *not* randomly chosen from a larger set of blocks. Then the treatment was randomized within each 'block'. Given this new information, I think it is okay to solve Kevin's question simply by using aov(Score.changes ~ Therapy + Block, data=table) assuming the fixed effects of 'block'. I would appreciate your correction if I am mistaken here. == T.K. (Tae-kyun) Kim Ph.D. student Department of Marketing Marshall School of Business University of Southern California == [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Displaying numerics to full double precision
On 12/5/07, Jeff Delmerico [EMAIL PROTECTED] wrote: Does anyone know if there is a way to change the display or storage settings so that the values will be displayed to their full precision? Or does rnorm only produce values to single precision? Any assistance would be greatly appreciated. As far as I know (beware, I'm a novice), internally R stores to and uses full precision. The display of the data, however, is controlled by digits. You'd need to put, say, options(digits=7) in your Rprofile.site (if it doesn't exist, creat it in the R etc/ folder). You might also be interested by scipen. Check ?options. Regards, Liviu __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Multiple stacked barplots on the same graph?
This command works: qplot(x=Categorie,y=Total,data=mydata,geom=bar,fill=Part) for your data. domenico vistocco Stéphane CRUVEILLER wrote: Hi, the same error message is displayed with geom=bar as parameter. here is the output of dput: dput(mydata) structure(list(Categorie = structure(c(1L, 12L, 8L, 2L, 5L, 7L, 16L, 6L, 15L, 11L, 10L, 13L, 14L, 3L, 4L, 9L, 17L, 1L, 12L, 8L, 2L, 5L, 7L, 16L, 6L, 15L, 11L, 10L, 13L, 14L, 3L, 4L, 9L, 17L ), .Label = c(Amino acid biosynthesis, Biosynthesis of cofactors, prosthetic groups, and carriers, Cell envelope, Cellular processes, Central intermediary metabolism, DNA metabolism, Energy metabolism, Fatty acid and phospholipid metabolism, Mobile and extrachromosomal element functions, Protein fate, Protein synthesis, Purines, pyrimidines, nucleosides, and nucleotides, Regulatory functions, Signal transduction, Transcription, Transport and binding proteins, Unknown function), class = factor), Part = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c(common, specific), class = factor), Total = c(3.03, 1.65, 1.52, 2.85, 3.4, 11.81, 10.51, 1.95, 2.08, 2.51, 2.23, 7.63, 1.88, 2.76, 7.21, 1.08, 20.75, 0.35, 0.17, 0.08, 0.18, 0.42, 2.05, 1.98, 0.63, 0.17, 0.2, 0.3, 1.58, 0.27, 0.83, 1.38, 3.56, 11.63), chr1 = c(4.55, 2.37, 1.77, 4.68, 3.19, 12.49, 13.56, 2.81, 3.13, 4.58, 3.26, 7.3, 2.06, 3.41, 7.9, 0.22, 22.45, 0.16, 0.06, 0.09, 0.19, 0.09, 0.7, 0.85, 0.22, 0.06, 0.03, 0.32, 0.66, 0.06, 0.63, 0.38, 1.14, 6.17), chr2 = c(1.68, 1.06, 1.55, 1.02, 4.57, 13.87, 7.85, 0.98, 1.06, 0.27, 1.2, 9.88, 2.13, 2.53, 7.71, 0.4, 22.38, 0.71, 0.35, 0.09, 0.22, 0.98, 3.9, 3.24, 0.22, 0.22, 0.49, 0.31, 2.79, 0.62, 1.33, 1.95, 0.44, 16), pl = c(0, 0, 0, 0, 0, 0.17, 4.27, 1.03, 0.34, 0, 0.68, 0.68, 0, 0.17, 1.54, 8.38, 5.3, 0, 0, 0, 0, 0, 2.22, 3.25, 4.44, 0.51, 0, 0.17, 1.88, 0, 0, 4.62, 28.72, 24.27)), .Names = c(Categorie, Part, Total, chr1, chr2, pl), class = data.frame, row.names = c(NA, -34L)) thx, Stéphane. hadley wickham wrote: On Dec 4, 2007 10:34 AM, Stéphane CRUVEILLER [EMAIL PROTECTED] wrote: Hi, I tried this method but it seems that there is something wrong with my data frame: when I type in: qplot(x=as.factor(Categorie),y=Total,data=mydata) It displays a graph with 2 points in each category... but if I add the parameter geom=histogram qplot(x=as.factor(Categorie),y=Total,data=mydata,geom=histogram) Error in storage.mode(test) - logical : object y not found any hint about this... Could you copy and paste the output of dput(mydata) ? (And I'd probably write the plot call as: qplot(Categorie, Total, data=mydata, geom=bar), since it is a bar plot, not a histogram) __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Significance of clarkevans in spatstat
Dear R-Users, I was wondering if there is a way to test the significance of the clarkevans statistic in spatstat package? I did not find any related function or the related values to calculate it by hand. does someone has any ideas? thank you, Jens -- + Dipl.Biol. Jens Oldeland University of Hamburg Biocentre Klein Flottbek and Botanical Garden Ohnhorststr. 18 22609 Hamburg, Germany Tel:0049-(0)40-42816-407 Fax:0049-(0)40-42816-543 Mail: [EMAIL PROTECTED] [EMAIL PROTECTED] (for attachments 2mb!!) http://www.biologie.uni-hamburg.de/bzf/fbda005/fbda005.htm + __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] weighted Cox proportional hazards regression
I'm getting unexpected results from the coxph function when using weights from counter-matching. For example, the following code produces a parameter estimate of -1.59 where I expect 0.63: I agree with Thomas' answer wrt using offset instead of weight. One way to understand this is to look at the score equation for the Cox model, which is sum over the deaths of (x[i] - xbar[i]) x[i] is the covariate vector of the ith death xbar[i] is the average of all the subjects who were at risk at the time of the ith death. In situations where one samples selected controls, the score equation will be correct if one fixes up xbar so that it is an estimate of the population mean (all those in the population that were at risk for a death) rather than being the mean of just those in the sample. Use of an offset statement allows one to reweight xbar without changing the rest of the score equation. It's kind of a trick, see Therneau and Li, Lifetime Data Analysis, 1999, p99-112 for a simple example of how it works. Langholz and Borgan give details on exactly how to correctly reweight using some old results from sampling theory - it is just a little bit more subtle than one would guess, but not too different from the obvious. Terry Therneau __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] rgdal for R.2.4.0?
On Wed, 5 Dec 2007, bernardo lagos alvarez wrote: Hi, Know anyone where to find the package rgdal for R.2.4.0? On CRAN: the current version has Package: rgdal Title: Bindings for the Geospatial Data Abstraction Library Version: 0.5-20 Date: 2007-11-07 Depends: R (= 2.3.0), methods, sp Or were you looking for a binary version for an unstated platform? (A Windows binary is there and should be available via the Rgui menus.) -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Junk or not Junk ???
Thank you As per advice from several R users I have set r-project.org, stat.math.ethz.ch,fhcrc.org, stat.ethz.ch, math.ethz.ch, hypatia.math.ethz.ch all to be safe domains But still some R emails go to Junk and require to be found manually I have explored the issue with Univ Wash computing to no avail Is this just how it is or have I still missed the fix to keep R emails out of junk? Thank you Loren Engrav Univ Wash Seattle From: Duncan Murdoch [EMAIL PROTECTED] Date: Mon, 03 Dec 2007 22:10:18 -0500 To: Loren Engrav [EMAIL PROTECTED] Cc: RHelp r-help@r-project.org Subject: Re: [R] Junk or not Junk On 03/12/2007 8:56 PM, Loren Engrav wrote: So a message from Benilton Carvalho [EMAIL PROTECTED] (sent by [EMAIL PROTECTED]) arrives and goes in the Junk Mail even tho I have set @r-project.org to not be junk Why does this go in Junk mail if @r-project.org is defined as not junk? Why are you asking us about how you have your mail filters set up? If you didn't set them up yourself, you should find out from your local admin who did, and ask them. Duncan Murdoch __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] confidence intervals for y predicted in non linear regression
Hi, Salut, You should use the package nsl2 (only for Linux distribution) Vous pouvez utiliser le package nls2 (Linux seulement) Regards, Souleymane Date: Tue, 4 Dec 2007 16:07:57 +0100 From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] CC: [EMAIL PROTECTED] Subject: Re: [R] confidence intervals for y predicted in non linear regression hi, hi all, you can consult these links: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/43008.html https://stat.ethz.ch/pipermail/r-help/2004-October/058703.html hope this help pierre Selon Florencio González [EMAIL PROTECTED]:Hi, I´m trying to plot a nonlinear regresion with the confidence bands for the curve obtained, similar to what nlintool or nlpredci functions in Matlab does, but I no figure how to. In nls the option is there but not implemented yet. Is there a plan to implement the in a relative near future? Thanks in advance, Florencio La información contenida en este e-mail y sus ficheros adjuntos es totalmente confidencial y no debería ser usado si no fuera usted alguno de los destinatarios. Si ha recibido este e-mail por error, por favor avise al remitente y bórrelo de su buzón o de cualquier otro medio de almacenamiento. This email is confidential and should not be used by anyone who is not the original intended recipient. If you have received this e-mail in error please inform the sender and delete it from your mailbox or any other storage mechanism. [[alternative HTML version deleted]]__ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Vous êtes plutôt Desperate ou LOST ? Personnalisez votre PC avec votre [[replacing trailing spam]] [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Quadratic programming
G'day Serge, On Wed, 5 Dec 2007 11:25:41 +0100 de Gosson de Varennes Serge (4100) [EMAIL PROTECTED] wrote: I am using the quadprog package and its solve.QP routine to solve and quadratic programming problem with inconsistent constraints, which obviously doesn't work since the constraint matrix doesn't have full rank. I guess it will help to fix some terminology first. In my book, inconsistent constraints are constraints that cannot be fulfilled simultaneously, e.g. something like x_1 = 3 and x_1 = 5 for an obvious example. Thus, a problem with inconsistent constraints cannot be solved, regardless of the rank of the constraint matrix. (Anyway, that matrix is typically not square, so would be be talking about full column rank or full row rank?) Of course, it can happen that the constraints are consistent but that there are some redundancy in the specified constraints, e.g. a simply case would be x_1 = 0, x_2 = 0 and x_1+x_2 = 0; if the first two constraints are fulfilled, then the last one is automatically fulfilled too. In my experience, it can happen that solve.QP comes to the conclusion that a constraint that ought to be fulfilled, given the constraints that have already been enforced, is deemed to be violated and to be inconsistent with the constraints already enforced. In that case solve.QP stops, rather misleadingly, with the message that the constraints are inconsistent. I guess the package should be worked over by someone with a better understanding of the kind of fudges that do not come back to bite and of finite precision arithmetic than the original author's appreciation of such issues when the code was written. ;-)) A way to solve this is to perturb the objective function and the constraints with a parameter that changes at each iteration (so you can dismiss it), but that's where it gets tricky! Solve.QP doesn't seem to admitt constant terms, it need Dmat (a matrix) and dvec (a vector) as defined in the package description. Now, some might object that a constant is a vector but the problem looks like this Min f(x) = (1/2)x^t Q x + D^t x + d It is a bit unclear to me what you call the constant term. Is it `d'? In that case, it does not perturb the constraints and it is irrelevant for the minimizer of f(x); also for the minimizer of f(x) under linear constraints. Regardless of d, the solution is always the same. I do not know of any quadratic programming solver that allows `d' as input, probably because it is irrelevant for determining the solution of the problem. Can anyone help me, PLEASEEE? In my experience, rescaling the problem might help, i.e. use Q* = Q/2 and D*=D/2 instead of the original Q and D; but do not forget to rescale the constraints accordingly. Or you might want to try another quadratic program solver in R, e.g. ipop() in package kernlab. Hope this helps. Best wishes, Berwin === Full address = Berwin A TurlachTel.: +65 6516 4416 (secr) Dept of Statistics and Applied Probability+65 6516 6650 (self) Faculty of Science FAX : +65 6872 3919 National University of Singapore 6 Science Drive 2, Blk S16, Level 7 e-mail: [EMAIL PROTECTED] Singapore 117546http://www.stat.nus.edu.sg/~statba __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] bootstrapping on the growth curve
Hi, I am trying to get 95% CI s around a quantile growth curve, but my result looks strange to me. Here is the function I defined myself boot.qregress-function(mat1, group, quantile, int, seed.1){ boot.fit-NULL set.seed(seed.1) for (i in 1:int){ index-sample((unique(mat1$Subject[mat1$Group==group])), length (unique(mat1$Subject[mat1$Group==group])), replace=TRUE) #make the bootstrapping dataset mat.junk-NULL for (j in 1: length(index)){ mat.junk-rbind(mat.junk, mat1[mat1$Subject==index[j], ]) } boot.fit-cbind(boot.fit, cobs(mat.junk$Day, mat.junk$Weight, constraint=none, degree=2, tau=quantile, lambda=-1)$fitted) } boot.fit } The curves I made from the bootstrapping is attached, I don't understand why for a group, the 5% curve drops suddenly around time 130. I am thinking about missingness since before 130 there are 50 patients, but after day 130 there are only 40 patients for this group. Any suggestions on the R-code (especially about how to do the bootstrapping for the growth curves) or why the drops happened would be appreciated. Thanks a lot, Suyan  __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
but unfortunately it did not work. What did not work? Provide some information more... By the way, isn't 'gfortran' the new GNU fortran compiler which replaced 'g77'? Or not on Ubuntu? I immediately realize that an amd64 is not the right processor to make life easy. ??? Example of bad extrapolation ;) Should I install the i386 version of R ? I assume you are talking about installing from source. You installed the i386 version of Ubuntu? Then yes. Else no. But rather let apt-get do it for you. Just install the binary provided by the Ubuntu community. It is the best way to get things working and avoid problems. Should I install another flavour of Linux ? It depends. Ubuntu is good to start, and has the widest users base; Archlinux my best choice (but you need to be already somewhat advanced). ScionForbai __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] about color palettes, colorRamp etc
[if you get this twice: it seems to have not made it through, yesterday] Earl == Earl F Glynn [EMAIL PROTECTED] on Mon, 3 Dec 2007 13:26:11 -0600 writes: Earl affy snp [EMAIL PROTECTED] wrote in message Earl news:[EMAIL PROTECTED] For example, it should go from very red---red---less red---darkgreen---very green coinciding with the descending order of values, just like the very left panel shown in http://www.bme.unc.edu/research/Bioinformatics.FunctionalGenomics.html Earl This looks like the MatLab palette that's in Earl tim.colors: Earl library(fields) # tim.colors: Matlab-like color palette Earl N - 100 Earl par(lend=square) Earl plot(rep(1,N), type=h, col=tim.colors(N), lwd=6, ylim=c(0,1)) Well, the R help page ?colorRamp in its 'examples' section has an example of this Matlab-lik color scheme, calling them 'jet.colors', easily constructed with the nice colorRampPalette() function [I've just posted about to R-help as well]. Please say example(colorRamp) in R and slowly watch the output, and I expect you will never ever want to use the horrible Matlab-like color palette again.. Regards, Martin Maechler, ETH Zurich Earl efg Earl Earl F. Glynn Earl Scientific Programmer Earl Stowers Institute for Medical Research __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] newbie lapply question
On Wed, 5 Dec 2007, Ranjan Bagchi wrote: Hi -- I just noticed the following (R 2.6.1 on OSX) lapply(c(as.Date('2007-01-01')), I) [[1]] [1] 13514 This is a bit surprising.. Why does lapply unclass the object? Sorry for such a basic question, I don't seem able to produce the right google keywords. Did you not read the help page?: Arguments: X: a vector (atomic or list) or an expressions vector. Other objects (including classed objects) will be coerced by 'as.list'. and as.list(c(as.Date('2007-01-01'))) [[1]] [1] 13514 BTW, the c() is redundant here: you are concatenating one item only. As to why as.list() removes the class, read its help page which tells you. Perhaps if you told us what you are trying to achieve we might be able to help you achieve 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Export to LaTeX using xtable() - Control the digits to the right of the separator [solved]
Hello everyone, The thread title speaks for itself. Here's the code that worked for me: numSummary(finance[,Employees], statistics=c(mean, sd, quantiles)) mean sd 0% 25% 50% 75% 100% n NA 11492.92 29373.14 1777 3040 4267 6553 179774 53 5 str(numSummary(finance[,Employees], statistics=c(mean, sd, quantiles))) List of 5 $ type : num 3 $ table : num [1, 1:7] 11493 29373 1777 3040 4267 ... ..- attr(*, dimnames)=List of 2 .. ..$ : chr .. ..$ : chr [1:7] mean sd 0% 25% ... $ statistics: chr [1:3] mean sd quantiles $ n : Named num 53 ..- attr(*, names)= chr data $ NAs : Named num 5 ..- attr(*, names)= chr data - attr(*, class)= chr numSummary xtable(numSummary(finance[,Employees], statistics=c(mean, sd, quantiles))$table, digit = c(0,0,2,2,2,0,0,0)) % latex table generated in R 2.6.1 by xtable 1.5-2 package % Wed Dec 5 14:37:51 2007 \begin{table}[ht] \begin{center} \begin{tabular}{} \hline mean sd 0\% 25\% 50\% 75\% 100\% \\ \hline 1 11493 29373.14 1777.00 3040.00 4267 6553 179774 \\ \hline \end{tabular} \end{center} \end{table} Regards, Liviu -- Forwarded message -- From: Romain Francois [EMAIL PROTECTED] Date: Dec 5, 2007 2:10 PM Subject: RE: [R] alternatives to latex() or xtable() ? To: Liviu Andronic [EMAIL PROTECTED] You need to look at the digits argument of xtable that would allow you to control this i think. xtable( numSummary( iris[,1:4] ) , digit = c( 0, 0, 2,2,2,2,2,2,0) ) % latex table generated in R 2.6.0 by xtable 1.5-2 package % Wed Dec 05 13:07:47 2007 \begin{table}[ht] \begin{center} \begin{tabular}{r} \hline mean sd 0\% 25\% 50\% 75\% 100\% n \\ \hline Sepal.Length 6 0.83 4.30 5.10 5.80 6.40 7.90 150 \\ Sepal.Width 3 0.44 2.00 2.80 3.00 3.30 4.40 150 \\ Petal.Length 4 1.77 1.00 1.60 4.35 5.10 6.90 150 \\ Petal.Width 1 0.76 0.10 0.30 1.30 1.80 2.50 150 \\ \hline \end{tabular} \end{center} \end{table} -Original Message- From: Liviu Andronic [mailto:[EMAIL PROTECTED] Sent: Wed 05/12/2007 13:07 To: Romain Francois Subject: Re: [R] alternatives to latex() or xtable() ? I have not yet understood how to set the number of displayed digits after the period (not sure how to express correctly in English) in the exported TeX code. For example, I would like to make all numbers display as integers. Or, I would like to have 123.00 numbers display as integers and the rest 123.212(3) display as 123.21. Do you know how this is done within R? (I understand that I can perfectly do this manually in the TeX code). Thanks in advance, Liviu __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Asymmetrically dependent variables to 2D-map?
Hello, I'm searching for a method which maps variables of this kind of table, see below, to 2-dimensional space, like in multidimensional scaling. However, this table is asymmetric: for example, variable T1 affects T2 more than T2 affects T1(0.41 vs. 0.21). DEPTABLE T1T2 T3T4 T1 0.00 0.41 0.24 1.18 T2 0.21 0.00 0.46 0.12 T3 0.80 0.89 0.00 0.20 T4 0.09 1.04 0.17 0.00 Any suggestions? Something like gplot+mds+weighted arrays? Atte Tenkanen University of Turku, Finland __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] 2/3d interpolation from a regular grid to another regular grid
On 2007-December-04 , at 21:38 , Scionforbai wrote: - krigging in package fields, which also requires irregular spaced data That kriging requires irregularly spaced data sounds new to me ;) It cannot be, you misread something (I feel free to say that even if I never used that package). Of Krigging I only know the name and general intent so I gladly line up to your opinion. I just read the description in ?Krig in the package fields which says: Fits a surface to irregularly spaced data. But there are probably other Krigging methods I overlooked. It can be tricky doing kriging, though, if you're not comfortable with a little bit of geostatistics. You have to infer a variogram model for each data set; you possibly run into non-stationarity or anisotropy, which are indeed very well treated (maybe at best) by kriging in one of its forms, but ... it takes more than this list to help you then; basically kriging requires modelling, so it is often very difficult to set up an automatic procedure. I can reccomend kriging if the spatial variability of your data (compared to grid refinement) is quite important. This was the impression I had too: that Krigging is an art in itself and that it requires you to know much about your data. My problem is simpler: the variability is not very large between grid points (it is oceanic current velocity data so it is highly auto-correlated spatially) and I can get grids fine enough for variability to be low anyway. So it is really purely numerical. In other simple cases, a wheighted mean using the (squared) inverse of the distance as wheight and a spherical neighbourhood could be the simpliest way to perform the interpolation. Yes, that would be largely enough for me. I had C routines for 2D polynomial interpolation of a similar cases and low order polynomes gave good results. I just hoped that R had that already coded somewhere in an handy and generic function rather than having to recode it myself in a probably highly specialized and not reusable manner. Thank you very much for you answer and if someone knows a function doing what is described above, that would be terrific. JiHO --- http://jo.irisson.free.fr/ __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] predict error for survreg with natural splines
Charles C. Berry wrote: On Wed, 5 Dec 2007, Gad Abraham wrote: Hi, The following error looks like a bug to me but perhaps someone can shed light on it: library(splines) library(survival) s - survreg(Surv(futime, fustat) ~ ns(age, knots=c(50, 60)), data=ovarian) n - data.frame(age=rep(mean(ovarian$age), 10)) predict(s, newdata=n) Error in qr.default(t(const)) : NA/NaN/Inf in foreign function call (arg 1) Thanks, Gad Gad, I think I have it now. survreg does not automatically place the boundary knots in its $terms component. You can force this by hand: Thanks Chuck and Moshe, manually setting the boundary fixes the problem. Cheers, Gad -- Gad Abraham Department of Mathematics and Statistics The University of Melbourne Parkville 3010, Victoria, Australia email: [EMAIL PROTECTED] web: http://www.ms.unimelb.edu.au/~gabraham __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] prediction R-squared
Hi, I need to compute a prediction R-squared for a linear regression. I have figured out how to compute Allen's PRESS statistic (using the PRESS function in the MPV library), but also want to compute the R-squared that goes along with this statistic. I have read that this is computed like an adjusted R-squared, but using the same regressions that are used to compute the PRESS statistic. I am trying to duplicate the prediction R-squared that is computed in Minitab. Thanks for any help! Tom [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Displaying numerics to full double precision
I'm working on a shared library of C functions for use with R, and I want to create a matrix in R and pass it to the C routines. I know R computes and supposedly stores numerics in double precision, but when I create a matrix of random numerics using rnorm(), the values are displayed in single precision, and also exported in single precision when I pass them out to my C routines. An example is below: a - matrix(rnorm(16, mean=10, sd=4), nrow=4) a [,1] [,2] [,3] [,4] [1,] 14.907606 17.572872 19.708977 9.809943 [2,] 9.322041 13.624452 7.745254 7.596176 [3,] 10.642408 6.151546 9.937434 6.913875 [4,] 14.617647 5.577073 8.217559 12.115465 storage.mode(a) [1] double Does anyone know if there is a way to change the display or storage settings so that the values will be displayed to their full precision? Or does rnorm only produce values to single precision? Any assistance would be greatly appreciated. Thanks, Jeff Delmerico -- View this message in context: http://www.nabble.com/Displaying-numerics-to-full-double-precision-tf4950807.html#a14175334 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] HTML help search in R 2.6.0 v 2.6.1
Michael Bibo wrote: I am running R on a corporate Windows XP SP2 machine on which I do not have administrator privileges or access to most settings in Control Panel. R is installed from my limited user account. The version of the JVM I have installed is perhaps best described as antique: system(paste(java -version),show.output.on.console=T) java version 1.4.1 Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.1-b21) Java HotSpot(TM) Client VM (build 1.4.1-b21, mixed mode) The HTML help search applet has worked for all versions of R up to and including 2.6.0, but in 2.6.1 the java applet is not initialised (Applet SearchEngine notinited). I have checked Appendix D of the admin installation manual, and the test java applet referred to does not load, but the web page says that java 1.4.2 is required. Other java applets do run in the browser, including R 2.6.0 HTML help search, so I presume java is enabled. Has something changed from 2.6.0 to 2.6.1 that may require JVM 1.4.1? If so, I can use that information to request an upgrade of my JVM. Hmm, could be. They got rebuilt on my system and committted at some point in the 2.6.1 run-in. and that system has 1.5.0. I did check that things still worked, but I didn't think the version would matter. The actual Java code is unchanged, so you could copy the .class files over from 2.6.0 ( let us know if it works) -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] correlation coefficient from qq plot
Hi, I am trying to figure out how to get the correlation coefficient for a QQ plot (residual plot). So to be more precise, I am creating the plot like this: qq.plot(rstudent(regrname), main = rformula, col=1) But want to also access (or compute) the correlation coefficient for that plot. Thanks, Tom [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] alternatives to latex() or xtable() ?
On 12/5/07, Romain Francois [EMAIL PROTECTED] wrote: Hello, My guess is that you are actually talking about the numSummary function in Rcmdr, not in abind. In that case, you can look how the structure of the output is like: str( numSummary( iris[,1:4] ) ) List of 4 $ type : num 3 $ table : num [1:4, 1:7] 5.843 3.057 3.758 1.199 0.828 ... ..- attr(*, dimnames)=List of 2 .. ..$ : chr [1:4] Sepal.Length Sepal.Width Petal.Length Petal.Width .. ..$ : chr [1:7] mean sd 0% 25% ... $ statistics: chr [1:3] mean sd quantiles $ n : Named num [1:4] 150 150 150 150 ..- attr(*, names)= chr [1:4] Sepal.Length Sepal.Width Petal.Length Petal.Width - attr(*, class)= chr numSummary and then use the table element from it: xtable( numSummary( iris[,1:4] )$table ) % latex table generated in R 2.6.0 by xtable 1.5-2 package % Wed Dec 05 12:16:44 2007 \begin{table}[ht] \begin{center} \begin{tabular}{} \hline mean sd 0\% 25\% 50\% 75\% 100\% \\ \hline Sepal.Length 5.84 0.83 4.30 5.10 5.80 6.40 7.90 \\ Sepal.Width 3.06 0.44 2.00 2.80 3.00 3.30 4.40 \\ Petal.Length 3.76 1.77 1.00 1.60 4.35 5.10 6.90 \\ Petal.Width 1.20 0.76 0.10 0.30 1.30 1.80 2.50 \\ \hline \end{tabular} \end{center} \end{table} Otherwise, you can define your own xtable.numSummary function that would wrap this up. (This one does not do everything as it does not take into account the groups argument of numSummary, so you might want to do something else if you have used it, ...) xtable.numSummary - function( x, ...){ + out - cbind( x$table, n = x$n ) + xtable( out, ... ) + } xtable( numSummary( iris[,1:4] ) ) % latex table generated in R 2.6.0 by xtable 1.5-2 package % Wed Dec 05 12:20:13 2007 \begin{table}[ht] \begin{center} \begin{tabular}{r} \hline mean sd 0\% 25\% 50\% 75\% 100\% n \\ \hline Sepal.Length 5.84 0.83 4.30 5.10 5.80 6.40 7.90 150.00 \\ Sepal.Width 3.06 0.44 2.00 2.80 3.00 3.30 4.40 150.00 \\ Petal.Length 3.76 1.77 1.00 1.60 4.35 5.10 6.90 150.00 \\ Petal.Width 1.20 0.76 0.10 0.30 1.30 1.80 2.50 150.00 \\ \hline \end{tabular} \end{center} \end{table} Hope this helps, Romain It helped. Thanks. Liviu __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Dealing with NA's in a data matrix
Henrique Dallazuanna wwwhsd at gmail.com writes: x[is.na(x)] - 0 On 05/12/2007, Amit Patel amitpatel_ak at yahoo.co.uk wrote: Hi I have a matrix with NA value that I would like to convert these to a value of 0. any suggestions also library(gdata) x - matrix(rnorm(16), nrow=4, ncol=4) x[1, 1] - NA NAToUnknown(x, unknown=0) Gregor __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] confidence intervals for y predicted in non linearregression
Hi Thanks for your suggestion, I'm trying to install this package in Ubuntu (7.10) but unsuccessfully. Also tried in MacOSX, and no success too. _ De: Ndoye Souleymane [mailto:[EMAIL PROTECTED] Enviado el: miércoles, 05 de diciembre de 2007 13:38 Para: [EMAIL PROTECTED]; Florencio González CC: [EMAIL PROTECTED] Asunto: RE: [R] confidence intervals for y predicted in non linearregression Hi, Salut, You should use the package nsl2 (only for Linux distribution) Vous pouvez utiliser le package nls2 (Linux seulement) Regards, Souleymane Date: Tue, 4 Dec 2007 16:07:57 +0100 From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] CC: [EMAIL PROTECTED] Subject: Re: [R] confidence intervals for y predicted in non linear regression hi, hi all, you can consult these links: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/43008.html https://stat.ethz.ch/pipermail/r-help/2004-October/058703.html hope this help pierre Selon Florencio González [EMAIL PROTECTED]: Hi, I´m trying to plot a nonlinear regresion with the confidence bands for the curve obtained, similar to what nlintool or nlpredci functions in Matlab does, but I no figure how to. In nls the option is there but not implemented yet. Is there a plan to implement the in a relative near future? Thanks in advance, Florencio La información contenida en este e-mail y sus ficheros adjuntos es totalmente confidencial y no debería ser usado si no fuera usted alguno de los destinatarios. Si ha recibido este e-mail por error, por favor avise al remitente y bórrelo de su buzón o de cualquier otro medio de almacenamiento. This email is confidential and should not be used by anyone who is not the original intended recipient. If you have received this e-mail in error please inform the sender and delete it from your mailbox or any other storage mechanism. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Besoin d'un e-mail ? Créez gratuitement un compte Windows Live Hotmail et bénéficiez d'un filtre antivirus gratuit ! Windows Live http://www.windowslive.fr/hotmail/default.asp Hotmail La información contenida en este e-mail y sus ficheros adjuntos es totalmente confidencial y no debería ser usado si no fuera usted alguno de los destinatarios. Si ha recibido este e-mail por error, por favor avise al remitente y bórrelo de su buzón o de cualquier otro medio de almacenamiento. This email is confidential and should not be used by anyone who is not the original intended recipient. If you have received this e-mail in error please inform the sender and delete it from your mailbox or any other storage mechanism. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] HTML help search in R 2.6.0 v 2.6.1
Peter Dalgaard p.dalgaard at biostat.ku.dk writes: Has something changed from 2.6.0 to 2.6.1 that may require JVM 1.4.1? If so, I can use that information to request an upgrade of my JVM. Hmm, could be. They got rebuilt on my system and committted at some point in the 2.6.1 run-in. and that system has 1.5.0. I did check that things still worked, but I didn't think the version would matter. The actual Java code is unchanged, so you could copy the .class files over from 2.6.0 ( let us know if it works) Excellent! That worked. Thanks, Peter. If this is going to be a continuing issue for future versions, would the best advice be to upgrade JVM anyway, at least to 1.5.0? Michael __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] File based configuration
See the following (the email seems to have wrapped 2 lines onto one at one point but it should be obvious): http://tolstoy.newcastle.edu.au/R/e2/help/07/06/18853.html On Dec 5, 2007 5:05 PM, Thomas Allen [EMAIL PROTECTED] wrote: I'm wanting to run R scripts non-interactively as part of a technology independent framework. I want control over the behaviour of these processes by specifying various global variables in a configuration file that would be passed as a command line argument. I'm wondering if you know of any R support for configuration file formats. (i.e. any functions that would read a configuration file of some common format) For example: -The .properties configuration format for java seems to be quite popular, would I have to read it in by writing some kind of java extension to R? -An XML configuration format could also be possible, but it's overkill for my needs. Any help would be greatly appreciated __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Information criteria for kmeans
Hello, how is, for example, the Schwarz criterion is defined for kmeans? It should be something like: k - 2 vars - 4 nobs - 100 dat - rbind(matrix(rnorm(nobs, sd = 0.3), ncol = vars), matrix(rnorm(nobs, mean = 1, sd = 0.3), ncol = vars)) colnames(dat) - paste(var,1:4) (cl - kmeans(dat, k)) schwarz - sum(cl$withinss)+ vars*k*log(nobs) Thanks for your help, Serguei Austrian Institute of Economic Research (WIFO) P.O.Box 91 Tel.: +43-1-7982601-231 1103 Vienna, AustriaFax: +43-1-7989386 Mail: [EMAIL PROTECTED] http://www.wifo.ac.at/Serguei.Kaniovski [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] nearest correlation to polychoric
Dear Jens, I've submitted a new version (0.7-4) of the polycor package to CRAN. The hetcor() function now uses your nearcor() in sfsmisc to make the returned correlation matrix positive-definite if it is not already. I know that quite some time has elapsed since you raised this issue, and I apologize for taking so long to deal with it. (I've also kept track of your suggestions for the sem package, and will respond to them when I next make substantial modifications to the package -- though not in the near future.) Thank you, John John Fox, Professor Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: Jens Oehlschlägel [mailto:[EMAIL PROTECTED] Sent: Friday, July 13, 2007 2:42 PM To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: RE: [R] nearest correlation to polychoric Dimitris, Thanks a lot for the quick response with the pointer to posdefify. Using its logic as an afterburner to the algorithm of Higham seems to work. Martin, Jens, could you make your code (mentioned below) available to the community, or even donate to be included as a new method of posdefify() ? Nice opportunity to give-back. Below is the R code for nearcor and .Rd help file. A quite natural place for nearcor would be John Fox' package polycor, what do you think? John? Best regards Jens Oehlschlägel # Copyright (2007) Jens Oehlschlägel # GPL licence, no warranty, use at your own risk #! \name{nearcor} #! \alias{nearcor} #! \title{ function to find the nearest proper correlation matrix given an improper one } #! \description{ #! This function smooths a improper correlation matrix as it can result from \code{\link{cor}} with \code{use=pairwise.complete.obs} or \code{\link[polycor]{hetcor}}. #! } #! \usage{ #! nearcor(R, eig.tol = 1e-06, conv.tol = 1e-07, posd.tol = 1e-08, maxits = 100, verbose = FALSE) #! } #! \arguments{ #! \item{R}{ a square symmetric approximate correlation matrix } #! \item{eig.tol}{ defines relative positiveness of eigenvalues compared to largest, default=1.0e-6 } #! \item{conv.tol}{ convergence tolerance for algorithm, default=1.0e-7 } #! \item{posd.tol}{ tolerance for enforcing positive definiteness, default=1.0e-8 } #! \item{maxits}{ maximum number of iterations allowed } #! \item{verbose}{ set to TRUE to verbose convergence } #! } #! \details{ #! This implements the algorithm of Higham (2002), then forces symmetry, then forces positive definiteness using code from \code{\link[sfsmisc]{posdefify}}. #! This implementation does not make use of direct LAPACK access for tuning purposes as in the MATLAB code of Lucas (2001). #! The algorithm of Knol DL and ten Berge (1989) (not implemented here) is more general in (1) that it allows contraints to fix some rows (and columns) of the matrix and (2) to force the smallest eigenvalue to have a certain value. #! } #! \value{ #! A LIST, with components #! \item{cor}{resulting correlation matrix} #! \item{fnorm}{Froebenius norm of difference of input and output} #! \item{iterations}{number of iterations used} #! \item{converged}{logical} #! } #! \references{ #!Knol, DL and ten Berge, JMF (1989). Least-squares approximation of an improper correlation matrix by a proper one. Psychometrika, 54, 53-61. #! \cr Higham (2002). Computing the nearest correlation matrix - a problem from finance, IMA Journal of Numerical Analysis, 22, 329-343. #! \cr Lucas (2001). Computing nearest covariance and correlation matrices. A thesis submitted to the University of Manchester for the degree of Master of Science in the Faculty of Science and Engeneering. #! } #! \author{ Jens Oehlschlägel } #! \seealso{ \code{\link[polycor]{hetcor}}, \code{\link{eigen}}, \code{\link[sfsmisc]{posdefify}} } #! \examples{ #! cat(pr is the example matrix used in Knol DL, ten Berge (1989)\n) #! pr - structure(c(1, 0.477, 0.644, 0.478, 0.651, 0.826, 0.477, 1, 0.516, #! 0.233, 0.682, 0.75, 0.644, 0.516, 1, 0.599, 0.581, 0.742, 0.478, #! 0.233, 0.599, 1, 0.741, 0.8, 0.651, 0.682, 0.581, 0.741, 1, 0.798, #! 0.826, 0.75, 0.742, 0.8, 0.798, 1), .Dim = c(6, 6)) #! #! nr - nearcor(pr)$cor #! plot(pr[lower.tri(pr)],nr[lower.tri(nr)]) #! round(cbind(eigen(pr)$values, eigen(nr)$values), 8) #! #! cat(The following will fail:\n) #! try(factanal(cov=pr, factors=2)) #! cat(and this should work\n) #! try(factanal(cov=nr, factors=2)) #! #! \dontrun{ #! library(polycor) #! #! n - 400 #! x - rnorm(n) #! y - rnorm(n) #! #! x1 - (x + rnorm(n))/2 #! x2 - (x + rnorm(n))/2 #! x3 - (x + rnorm(n))/2 #! x4 - (x + rnorm(n))/2 #! #! y1 - (y + rnorm(n))/2 #!
[R] Plotting error bars in xy-direction
Dear R-help, I am looking for a function that will plot error bars in x- or y-direction (or both), the same as the Gnuplot function 'plot' can achieve with: plot file.dat with xyerrorbars,... Rsite-searching led me to the functions 'errbar' and 'plotCI' in the Hmisc, gregmisc, and plotrix packages. As I understand the descriptions and examples, none of these functions provides horizontal error bars. Looking into 'errbar' and using segments, I wrote a small function for myself adding these kinds of error bars to existing plots. I would still be interested to know what the standard R solution is. Regards, Hans Werner __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] confidence intervals for y predicted in non linear regression
I don't think this is referring to the nls2 package on CRAN but rather something else. On Dec 5, 2007 7:37 AM, Ndoye Souleymane [EMAIL PROTECTED] wrote: Hi, Salut, You should use the package nsl2 (only for Linux distribution) Vous pouvez utiliser le package nls2 (Linux seulement) Regards, Souleymane Date: Tue, 4 Dec 2007 16:07:57 +0100 From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] CC: [EMAIL PROTECTED] Subject: Re: [R] confidence intervals for y predicted in non linear regression hi, hi all, you can consult these links: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/43008.html https://stat.ethz.ch/pipermail/r-help/2004-October/058703.html hope this help pierre Selon Florencio González [EMAIL PROTECTED]: Hi, I´m trying to plot a nonlinear regresion with the confidence bands for the curve obtained, similar to what nlintool or nlpredci functions in Matlab does, but I no figure how to. In nls the option is there but not implemented yet. Is there a plan to implement the in a relative near future? Thanks in advance, Florencio La información contenida en este e-mail y sus ficheros adjuntos es totalmente confidencial y no debería ser usado si no fuera usted alguno de los destinatarios. Si ha recibido este e-mail por error, por favor avise al remitente y bórrelo de su buzón o de cualquier otro medio de almacenamiento. This email is confidential and should not be used by anyone who is not the original intended recipient. If you have received this e-mail in error please inform the sender and delete it from your mailbox or any other storage mechanism. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. _ Vous êtes plutôt Desperate ou LOST ? Personnalisez votre PC avec votre [[replacing trailing spam]] [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] logistic regression using glm,which y is set to be 1
Dear friends : using the glm function and setting family=binomial, I got a list of coefficients. The coefficients reflect the effects of predicted variables on the probability of the response to be 1. My response variable consists of A and D . I don't know which level of the response was set to be 1. is the first element of the response set to be 1? Thank all in advance. Regards, - Best regards, Bin Yue * student for a Master program in South Botanical Garden , CAS -- View this message in context: http://www.nabble.com/logistic-regression-using-%22glm%22%2Cwhich-%22y%22-is-set-to-be-%221%22-tf4953617.html#a14185060 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] logistic regression using glm,which y is set to be 1
On Wed, 2007-12-05 at 18:06 -0800, Bin Yue wrote: Dear friends : using the glm function and setting family=binomial, I got a list of coefficients. The coefficients reflect the effects of predicted variables on the probability of the response to be 1. My response variable consists of A and D . I don't know which level of the response was set to be 1. is the first element of the response set to be 1? Thank all in advance. Regards, - Best regards, Bin Yue As per the Details section of ?glm: For binomial and quasibinomial families the response can also be specified as a factor (when the first level denotes failure and all others success) ... So use: levels(response.variable) and that will give you the factor levels, where the first level is 0 and the second level is 1. If you work in a typical English based locale with default alpha based level ordering, it will likely be A (Alive?) is 0 and D (Dead?) is 1. HTH, Marc Schwartz __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] logistic regression using glm,which y is set to be 1
Dear Marc Schwartz: When I ask R2.6.0 for windows, the information it gives does not contain much about family=binomial . You said that there is a detail section of ?glm. I want to read it thoroughly. Could you tell me where and how I can find the detail section of ?glm. Thank you very much . Best regards, Bin Yue Marc Schwartz wrote: On Wed, 2007-12-05 at 18:06 -0800, Bin Yue wrote: Dear friends : using the glm function and setting family=binomial, I got a list of coefficients. The coefficients reflect the effects of predicted variables on the probability of the response to be 1. My response variable consists of A and D . I don't know which level of the response was set to be 1. is the first element of the response set to be 1? Thank all in advance. Regards, - Best regards, Bin Yue As per the Details section of ?glm: For binomial and quasibinomial families the response can also be specified as a factor (when the first level denotes failure and all others success) ... So use: levels(response.variable) and that will give you the factor levels, where the first level is 0 and the second level is 1. If you work in a typical English based locale with default alpha based level ordering, it will likely be A (Alive?) is 0 and D (Dead?) is 1. HTH, Marc Schwartz __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. - Best regards, Bin Yue * student for a Master program in South Botanical Garden , CAS -- View this message in context: http://www.nabble.com/logistic-regression-using-%22glm%22%2Cwhich-%22y%22-is-set-to-be-%221%22-tf4953617.html#a14185819 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Working with ts objects
anscombe is built into R already so you don't need to read it in. An intercept is the default in lm so you don't have to specify it. opar - par(mfrow = c(2,2)) plot(y1 ~ x1, anscombe) reg - lm(y1 ~ x1, anscombe) reg abline(reg) ...etc... par(opar) Note that plot(anscombe[1:2]) and lm(anscombe[2:1]) also work. read.table returns a data frame whereas ts requires a vector or matrix so none of your ts code will work. as.matrix(DF) or data.matrix(DF) will convert data frame DF to a matrix. On Dec 5, 2007 2:30 PM, Richard Saba [EMAIL PROTECTED] wrote: I am relatively new to R and object oriented programming. I have relied on SAS for most of my data analysis. I teach an introductory undergraduate forecasting course using the Diebold text and I am considering using R in addition to SAS and Eviews in the course. I work primarily with univariate or multivariate time series data. I am having a great deal of difficulty understanding and working with ts objects particularly when it comes to referencing variables in plot commands or in formulas. The confusion is amplified when certain procedures (lm for example) coerce the ts object into a data.frame before application with the results that the output is stored in a data.frame object. For example the two sets of code below replicate examples from chapter 2 and 6 in the text. In the first set of code if I were to replace anscombe-read.table(fname, header=TRUE) with anscombe-ts(read.table(fname, header=TRUE)) the plot() commands would generate errors. The objects x1, y1 ... would not be recognized. In this case I would have to reference the specific column in the anscombe data set. If I would have constructed the data set from several different data sets using the ts.intersect() function (see second code below)the problem becomes even more involved and keeping track of which columns are associated with which variables can be rather daunting. All I wanted was to plot actual vs. predicted values of hstarts and the residuals from the model. Given the difficulties I have encountered I know my students will have similar problems. Is there a source other than the basic R manuals that I can consult and recommend to my students that will help get a handle on working with time series objects? I found the Shumway Time series analysis and its applications with R Examples website very helpful but many practical questions involving manipulation of time series data still remain. Any help will be appreciated. Thanks, Richard Saba Department of Economics Auburn University Email: [EMAIL PROTECTED] Phone: 334 844-2922 anscombe-read.table(fname, header=TRUE) names(anscombe)-c(x1,y1,x2,y2,x3,y3,x4,y4) reg1-lm(y1~1 + x1, data=anscombe) reg2-lm(y2~1 + x2, data=anscombe) reg3-lm(y3~1 + x3, data=anscombe) reg4-lm(y4~1 + x4, data=anscombe) summary(reg1) summary(reg2) summary(reg3) summary(reg4) par(mfrow=c(2,2)) plot(x1,y1) abline(reg1) plot(x2,y2) abline(reg2) plot(x3,y3) abline(reg3) plot(x4,y4) abline(reg4) .. fname-file.choose() tab6.1-ts(read.table(fname, header=TRUE),frequency=12,start=c(1946,1)) month-cycle(tab6.1) year-floor(time(tab6.1)) dat1-ts.intersect(year,month,tab6.1) dat2-window(dat1,start=c(1946,1),end=c(1993,12)) reg1-lm(tab6.1~1+factor(month),data=dat2, na.action=NULL) summary(reg1) hstarts-dat2[,3] plot1-ts.intersect(hstarts,reg1$fitted.value,reg1$resid) plot.ts(plot1[,1]) lines(plot1[,2], col=red) plot.ts(plot[,3], ylab=Residuals) __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] snow package on multi core unix box
Is the rmpi package (or rpvm) needed to exploit multiple cores on a single unix box using the snow package. The documentation of the package does not provide info about setting up a single machine with multiple cores. Also, if how effective is it to run a bayesian simulation on parallel (or distributed) processors using the snow package. Thanks, Saeed __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Segmented regression
Hello all, I have 3 time series (tt) that I've fitted segmented regression models to, with 3 breakpoints that are common to all, using code below (requires segmented package). However I wish to specifiy a zero coefficient, a priori, for the last segment of the KW series (green) only. Is this possible to do with segmented? If not, could someone point in a direction? The final goal is to compare breakpoint sets for differences from those derived from other data. Thanks in advance, Brendan. library(segmented) df-data.frame(y=c(0.12,0.12,0.11,0.19,0.27,0.28,0.35,0.38,0.46,0.51,0.5 8,0.59,0.60,0.57,0.64,0.68,0.72,0.73,0.78,0.84,0.85,0.83,0.86,0.88,0.88, 0.95,0.95,0.93,0.92,0.97,0.86,1.00,0.85,0.97,0.90,1.02,0.95,0.54,0.53,0. 50,0.60,0.70,0.74,0.78,0.82,0.88,0.83,1.00,0.85,0.96,0.84,0.86,0.82,0.86 ,0.84,0.84,0.84,0.77,0.69,0.61,0.67,0.73,0.65,0.55,0.58,0.56,0.60,0.50,0 .50,0.42,0.43,0.44,0.42,0.40,0.51,0.60,0.63,0.71,0.74,0.82,0.82,0.85,0.8 9,0.91,0.87,0.91,0.93,0.95,0.95,0.97,1.00,0.96,0.90,0.86,0.91,0.94,0.96, 0.88,0.88,0.88,0.92,0.82,0.85), tt=c(141.6,141.6,141.6,183.2,212.8,227.0,242.4,271.5,297.4,312.3,331.4,3 42.4,346.3,356.6,371.6,408.8,408.8,419.5,434.4,464.5,492.6,521.7,550.5,5 50.3,565.4,588.0,602.9,623.7,639.6,647.9,672.6,680.6,709.7,709.7,750.2,7 50.2,750.2,141.6,141.6,141.6,183.2,212.8,227.0,242.4,271.5,297.4,312.3,3 31.4,342.4,346.3,356.6,371.6,408.8,408.8,419.5,434.4,464.5,492.6,521.7,5 50.5,550.3,565.4,588.0,602.9,623.7,639.6,647.9,672.6,680.6,709.7,709.7,1 41.6,141.6,141.6,183.2,212.8,227.0,242.4,271.5,297.4,312.3,331.4,342.4,3 46.3,356.6,371.6,408.8,408.8,419.5,434.4,464.5,492.6,521.7,550.5,550.3,5 65.4,588.0,602.9,623.7,639.6,647.9,672.6,709.7), group=c(rep(RKW,37),rep(RWC,34),rep(RKV,32))) init.bp - c(297.4,639.6,680.6) lm.1 - lm(y~tt+group,data=df) seg.1 - segmented(lm.1, seg.Z=~tt, psi=list(tt=init.bp)) version _ platform i386-pc-mingw32 arch i386 os mingw32 system i386, mingw32 status major 2 minor 6.0 year 2007 month 10 day03 svn rev43063 language R version.string R version 2.6.0 (2007-10-03) DISCLAIMER**...{{dropped:15}} __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] hclust in heatmap.2
Check the Rowv, Colv options to heatmap.2 data(mtcars) x - as.matrix(mtcars) heatmap.2(x, Rowv=FALSE, dendrogram=column) -Ashoka Scientist - Pacific Northwest National Lab On Dec 5, 2007 4:20 PM, affy snp [EMAIL PROTECTED] wrote: Dear list, I am using heatmap.2(x) to draw a heatmap. Ideally, I want to the matrix x clustered only by columns and keep the original order of rows unchanged. Is there a way to do that in heatmap.2()? Thanks a lot! Any suggestions will be appreciated! Best, Allen [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] logistic regression using glm,which y is set to be 1
Dear all: By comparing glmresult$y and model.response(model.frame(glmresult)), I have found out which one is set to be TRUE and which FALSE.But it seems that to fit a logistic regression , logit (or logistic) transformation has to be done before regression. Does anybody know how to obtain the transformation result ? It is hard to settle down before knowing the actual process R works . I have read some books and the ?glm help file , but what they told me was not sufficient. Best wishes , Bin Yue Weiwei Shi wrote: Dear Bin: you type ?glm in R console and you will find the Detail section of help file for glm i pasted it for you too Details A typical predictor has the form response ~ terms where response is the (numeric) response vector and terms is a series of terms which specifies a linear predictor for response. For binomialand quasibinomial families the response can also be specified as a factorfile:///Library/Frameworks/R.framework/Versions/2.6/Resources/library/base/html/factor.html (when the first level denotes failure and all others success) or as a two-column matrix with the columns giving the numbers of successes and failures. A terms specification of the form first + second indicates all the terms in first together with all the terms in second with duplicates removed. The terms in the formula will be re-ordered so that main effects come first, followed by the interactions, all second-order, all third-order and so on: to avoid this pass a terms object as the formula. A specification of the form first:second indicates the the set of terms obtained by taking the interactions of all terms in first with all terms in second. The specification first*second indicates the *cross* of first and second. This is the same as first + second + first:second. glm.fit is the workhorse function. If more than one of etastart, start and mustart is specified, the first in the list will be used. It is often advisable to supply starting values for a quasifile:///Library/Frameworks/R.framework/Versions/2.6/Resources/library/stats/html/family.html family, and also for families with unusual links such as gaussian(log). All of weights, subset, offset, etastart and mustart are evaluated in the same way as variables in formula, that is first in data and then in the environment of formula. On Dec 5, 2007 10:41 PM, Bin Yue [EMAIL PROTECTED] wrote: Dear Marc Schwartz: When I ask R2.6.0 for windows, the information it gives does not contain much about family=binomial . You said that there is a detail section of ?glm. I want to read it thoroughly. Could you tell me where and how I can find the detail section of ?glm. Thank you very much . Best regards, Bin Yue Marc Schwartz wrote: On Wed, 2007-12-05 at 18:06 -0800, Bin Yue wrote: Dear friends : using the glm function and setting family=binomial, I got a list of coefficients. The coefficients reflect the effects of predicted variables on the probability of the response to be 1. My response variable consists of A and D . I don't know which level of the response was set to be 1. is the first element of the response set to be 1? Thank all in advance. Regards, - Best regards, Bin Yue As per the Details section of ?glm: For binomial and quasibinomial families the response can also be specified as a factor (when the first level denotes failure and all others success) ... So use: levels(response.variable) and that will give you the factor levels, where the first level is 0 and the second level is 1. If you work in a typical English based locale with default alpha based level ordering, it will likely be A (Alive?) is 0 and D (Dead?) is 1. HTH, Marc Schwartz __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. - Best regards, Bin Yue * student for a Master program in South Botanical Garden , CAS -- View this message in context: http://www.nabble.com/logistic-regression-using-%22glm%22%2Cwhich-%22y%22-is-set-to-be-%221%22-tf4953617.html#a14185819 Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Weiwei Shi, Ph.D Research Scientist GeneGO, Inc. Did you always know? No, I did not. But I believed... ---Matrix III [[alternative HTML version deleted]] __ R-help@r-project.org mailing list
Re: [R] Which Linux OS on Athlon amd64, to comfortably run R?
On Thu, 6 Dec 2007, Emmanuel Charpentier wrote: Prof Brian Ripley a écrit : Note that Ottorino has only 1GB of RAM installed, which makes a 64-bit version of R somewhat moot. See chapter 8 of http://cran.r-project.org/doc/manuals/R-admin.html Thank you for this reminder|tip ! I didn't re-read this document since ... oh, my ... very early (1.x ?) versions. At which time my favorite peeve against R was the fixed memory allocation scheme. I would have thought that 64 bits machines could take advantage of a wider bus and (marginally ?) faster instructions to balance larger pointers. Am I mistaken ? Yes, it is more complex than that. If you run 32-bit instructions on a x86_64, the physical bus is the same as when you run 64-bit instructions. The larger code usually means the CPU caches spill more often, and some 64-bit chips have more 32-bit than 64-bit registers which allows better scheduling. The R-admin manual reports on some empirical testing. But when you have limited RAM the larger code and data for a 64-bit build will cause more swapping and that is likely to dominate performance issues on large problems. Note that the comparisons depend on both the chip and the OS: it seems that on Mac OS 10.5 on a Core 2 Duo the 64-bit version is faster (on small examples). The original enquiry was about 'amd64 linux', but I've checked Intel Core 2 Duo as well: on my box 64-bit builds are faster than 32-bit ones, whereas the reverse is true for Opterons. So it seems that the architectural differences of Core 2 Duo vs AMD64 do affect the issue. -- 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@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Using expression in Hmisc Key()
Michael Kubovy kubovy at virginia.edu writes: Dear r-helpers, How do I tell xYplot() and Key() that I want the labels in italic? Key(x = 0.667, y = 0.833, other = list(title = expression(italic(v)), cex.title = 1, labels = c(expression(italic(b)), expression(italic(c)), expression(italic(d) dev.off() Michael, I have submit a similar case last week to the Bug tracker. Maybe you can raise that enhancement request to defect http://biostat.mc.vanderbilt.edu/trac/Hmisc/ticket/21 Dieter --- The Key function generated by some plot commands should have a ... parameter. Otherwise, the ... in rlegend is useless, and it would be nice to be able to suppress the box, for example. Key = function (x = NULL, y = NULL, lev = c(No Fail, Fail), pch = c(16, 1)) { .. part omitted rlegend(x, y, legend = lev, pch = pch, ...) invisible() } __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.