Re: [R] Equivalent of Excel pivot tables in R
See also the reshape package. Cheers, Simon. Simon Blomberg, BSc (Hons), PhD, MAppStat. Lecturer and Consultant Statistician Faculty of Biological and Chemical Sciences The University of Queensland St. Lucia Queensland 4072 Australia T: +61 7 3365 2506 email: S.Blomberg1_at_uq.edu.au Policies: 1. I will NOT analyse your data for you. 2. Your deadline is your problem. The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data. - John Tukey. -Original Message- From: [EMAIL PROTECTED] on behalf of ppaarrkk Sent: Sat 26/04/2008 7:54 AM To: r-help@r-project.org Subject: [R] Equivalent of Excel pivot tables in R Can somebody tell me how to do the equivalent of a pivot table in R ? For example, if I have : var1var2 var3 a x10 b y20 a z10 b z20 a z10 b z20 I could have : x y z a1 0 2 b0 1 2 where entries in the table are counts of var3. x y z a10 0 20 b0 20 40 where entries are sums of var3. I would expect it to be tapply(), but I can't see how it would be done. Any suggestions please. -- View this message in context: http://www.nabble.com/Equivalent-of-Excel-pivot-tables-in-R-tp16906289p16906289.html 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. [[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] Another Question Regarding the nnet package
I have one other question regarding the nnet package in R: What is the value of the step size (often referred to as eta) for the gradient descent function? I dont see anywhere in the API where this value can be modified. Thanks, -Colin __ 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] matrix from list
OL == Olivier Lefevre [EMAIL PROTECTED] on Sat, 26 Apr 2008 01:31:16 +0200 writes: OL Greg Snow wrote: The '[[' only returns a single element from a data structure OL I know but that is precisely what I find arbitrary. OL Anyway you are right that it would still return the kind of object, only OL subsetted, which is not I want. As someone kindly pointed out to me OL offline, besides the unlist trick posted before, you can also achieve the OL desired result with OL m - as.matrix(l) OL dimnames(m) - NULL Two remarks on these two lines: - Slightly better for the 99% user- case of coercing a 'data.frame' (the result of read.table() or similar) ato a matrix is m - data.matrix(.) The difference to as.matrix() is that data.matrix() also produces a numeric matrix in the case the data frame contains factors. Consider (dd - data.frame(x= rnorm(10), f = gl(2,5)) as.matrix(dd) ## not what you want data.matrix(dd) ## what you want - I don't understand why you want to throw away the dimnames of the resulting numeric matrix. Rather, in general you'd want to keep them. Martin Maechler, ETH Zurich __ 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] quasi-random sequences
Dear list useRs, I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, as in: N - 100 positions - matrix(rnorm(2 * N, mean = 0 , sd= 0.5), nrow=N) sizes-rnorm(N, mean = 0 , sd= 1) plot(positions,type=p,cex=sizes) My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, library(gsl) g - qrng_alloc(type=sobol,dim=2) qrng_get(g,n= N) -xy plot((xy),t=p,cex=0.5) but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. Thanks for any comment / suggestion, Baptiste _ Baptiste Auguié Physics Department University of Exeter Stocker Road, Exeter, Devon, EX4 4QL, UK Phone: +44 1392 264187 http://newton.ex.ac.uk/research/emag http://projects.ex.ac.uk/atto __ 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] quasi-random sequences
baptiste Auguié wrote: Dear list useRs, I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, as in: The quasi-random sequences are useful for integration, but they're not random in the sense of being unpredictable. Here's an idea for a sequence that might work for you: divide the square into an n x n grid of smaller squares. Randomly select a smaller square, then uniformly select a point within it. (Or just permute the list of all n^2 squares, and proceed through the permuted list.) This won't guarantee a separation, but it will tend to lead to one. If you want a guarantee, then only select from squares with even coordinates. Once you've selected all the squares in the grid, go to a 2n x 2n grid, and repeat. (If you are doing the even coordinates modification, you'll have had to select non-uniformly for this to work.) Duncan Murdoch N - 100 positions - matrix(rnorm(2 * N, mean = 0 , sd= 0.5), nrow=N) sizes-rnorm(N, mean = 0 , sd= 1) plot(positions,type=p,cex=sizes) My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, library(gsl) g - qrng_alloc(type=sobol,dim=2) qrng_get(g,n= N) -xy plot((xy),t=p,cex=0.5) but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. Thanks for any comment / suggestion, Baptiste _ Baptiste Auguié Physics Department University of Exeter Stocker Road, Exeter, Devon, EX4 4QL, UK Phone: +44 1392 264187 http://newton.ex.ac.uk/research/emag http://projects.ex.ac.uk/atto __ 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] quasi-random sequences
On Sat, 26 Apr 2008, baptiste Auguié wrote: Dear list useRs, I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, as in: N - 100 positions - matrix(rnorm(2 * N, mean = 0 , sd= 0.5), nrow=N) sizes-rnorm(N, mean = 0 , sd= 1) plot(positions,type=p,cex=sizes) My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, That is a hard-core point process, e.g. a special case of a Strauss process. You will find ways to simulate such a process (there are several processes, and several ways for most) in the various spatial packages, including 'spatial' itself. I think you have misunderstood the point of quasi-random sequences (which, given the exposition in the edition of NR I have, would be easy to do). library(gsl) g - qrng_alloc(type=sobol,dim=2) qrng_get(g,n= N) -xy plot((xy),t=p,cex=0.5) but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. Thanks for any comment / suggestion, Baptiste _ Baptiste Auguié Physics Department University of Exeter Stocker Road, Exeter, Devon, EX4 4QL, UK Phone: +44 1392 264187 http://newton.ex.ac.uk/research/emag http://projects.ex.ac.uk/atto __ 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. -- 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] Ubuntu vs Mac OS X
Three comments, all from the 'R Installation and Adminstration' manual. 1) In general there is a small performance penalty for using a 64-bit version of R (larger pointers to move around). 2) For speed, bulld yourself a non-shared-library version of R. Those binary maintainers are supplying an R-shlib version that suits their agendas but maybe not yours. You may also want to increase the optimization level on the compiler. 3) If you are doing matrix algebra you want an optimized BLAS. The CRAN build of R for Mac OS by default uses vecLib (which is optimized by Apple). On other OSes R can be built to use optimized BLAS (see the manual) -- my understanding is that on Debian derivatives it is built to make use of the system libblas, and you can install versions of the latter based on ATLAS. Some optimized BLAS can make use of multiple CPUs. You have not told us what chip your '3.2 GHz Dual-core' is -- it will matter for an optimized BLAS. On Core 2 Duos I tend to use the (academic use only) BLAS by Dr Goto. On my modest Intel E6600 Core 2 Duo I get system.time(coco2-eigen(coco)) user system elapsed 11.984 0.052 12.065 system.time(coco3-coco%*%coco) user system elapsed 1.694 0.000 1.694 with vanilla 64-bit R compiled at -O3 and system.time(coco2-eigen(coco)) user system elapsed 7.450 0.062 3.912 system.time(coco3-coco%*%coco) user system elapsed 0.249 0.007 0.132 with the Goto BLAS. These are using both cores, as you see from the elapsed times. It is the work of moments to swap the BLAS used, which is how I did that test. On Fri, 25 Apr 2008, Guillaume Blanchet wrote: Hi ! I have installed R-2.7.0 64-bits on a computer where the new Ubuntu version (8.04) was installed. This computer is a 3.2 GHz Dual-core with 2 Gb of RAM. To test how fast (I though !) this machine is, I compared two simple matrix manipulations between Ubuntu and my laptop (a MacBook Pro 2.2 GHz with also 2 Gb of RAM under Mac OS X 10.5.2). By the way, I install the normal version of R on my MacBook Pro (32-bits) I decided to use the following matrix manipulation because they are typical manipulation that I was planning to do repetitively and with much larger matrices for my PhD project. First matrix manipulation: coco-matrix(rnorm(100),1000,1000) system.time(coco2-eigen(coco)) Results on the Mac: user system elapsed 7.301 0.974 7.376 Results on Ubuntu: user system elapsed 20.573 0.088 20.766 Second matrix manipulation: coco-matrix(rnorm(100),1000,1000) system.time(coco3-coco%*%coco) Results on the Mac: user system elapsed 0.360 0.024 0.224 Results on Ubuntu: user system elapsed 2.756 0.000 2.758 If I understand these results right, my MacBook Pro is far better than this Desk computer. I know Macs are good but that difference of performance is surprising me, did I do something wrong considering the large difference in processor? Thanks in advance Guillaume Blanchet __ 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. -- 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] How to overlap two density plots?
Very näively, you could do something like this, plot(density(A)) lines(density(B),col=2) , and tinker your xlim and ylim as suitable. The Cairo library gives a pretty example, data(iris) attach(iris) plot(Petal.Length, rep(-0.03,length(Species)), xlim=c(1,7), ylim=c(0,1.7), xlab=Petal.Length, ylab=Density, pch=21, cex=1.5, col=#0001, main = Iris (yet again), bg=c(#ff20,#00ff0020,#ff20)[unclass(Species)]) for (i in 1:3) polygon(density(Petal.Length[unclass(Species)==i],bw=0.2),col=c(#ff40,#00ff0040,#ff40)[i]) __ 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 overlap two density plots?
My apologies, the previous replies did not show up for me. On Sat, Apr 26, 2008 at 1:49 PM, Jared O'Connell [EMAIL PROTECTED] wrote: Very näively, you could do something like this, plot(density(A)) lines(density(B),col=2) , and tinker your xlim and ylim as suitable. The Cairo library gives a pretty example, data(iris) attach(iris) plot(Petal.Length, rep(-0.03,length(Species)), xlim=c(1,7), ylim=c(0,1.7), xlab=Petal.Length, ylab=Density, pch=21, cex=1.5, col=#0001, main = Iris (yet again), bg=c(#ff20,#00ff0020,#ff20)[unclass(Species)]) for (i in 1:3) polygon(density(Petal.Length[unclass(Species)==i],bw=0.2),col=c(#ff40,#00ff0040,#ff40)[i]) __ 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] 3 questions: debug R script, multi-level sorts, and multi-gsub
Ng Stanley [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: 2) I have looked deeply into R, but can't fine multi-level sorts. At the top of the help page for sort... Sort (or order) a vector or factor (partially) into ascending (or descending) order. For ordering along more than one variable, e.g., for sorting data frames, see _order_. In the examples on the order page, it shows what I suspect you mean by a multilevel sort (although you offer no examples). (ii - order(x - c(1,1,3:1,1:4,3), y - c(9,9:1), z -c(2,1:9))) [1] 6 5 2 1 7 4 10 8 3 9 dd - transform(data.frame(x,y,z), + z = factor(z, labels=LETTERS[9:1])) dd x y z 1 1 9 H 2 1 9 I 3 3 8 H 4 2 7 G 5 1 6 F 6 1 5 E 7 2 4 D 8 3 3 C 9 4 2 B 10 3 1 A dd[ order(x, -y, z) ,] x y z 2 1 9 I 1 1 9 H 5 1 6 F 6 1 5 E 4 2 7 G 7 2 4 D 3 3 8 H 8 3 3 C 10 3 1 A 9 4 2 B -- David Winsemius 3) I have a vector of words, several of which need to be replaced by different unique words. I am using multi-lines of gsub, is there any elegant alternative ? [[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] Help with simulation of heteroskedasticity
Hello guys! Sorry to bother with such a question I was trying to generate a monte carlo simulation with heteroskedasticity errors. but I am not sure if the command line that I had wrote is quite correct. the type of heteroskedasticity that I want to create is such as var(e) = var(x^4) I began my work with this x- rnorm (100, 2,0.4) # generating an indepedent random variable e- rnorm(100,0,x^2) # generating the error term y.fitted- 0.5*(x) +3 y- y.fitted+e And then I wrote an for loop code which could give me the results of a t-test. It seems that the simulation worked good because the main result is that the t-test was no longer meanfull (as it is expected) But my main doubt is if this code was capable to generate the type of heteroskedasticity that I wanted. I tried to understand how the rnorm function generates the numbers but i could not grab much So could anyone help me telling if this kind of work is okay? or is there any moree inteligent way to do the job? thanks!!! [[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] sys.time question - how to improve R performance
Dear R-users, I run my program a difference between sys.times is as follows: user system elapsed 60167.53 2848.75 63278.93 I am quite puzzled how it may happen that system time is so much shorter than user time, I have 2 core computer most of the time (90%) I was not doing anything else with the computer. What could be the reason of such a difference? I switched off file indexing and all other resident programs.. My program include writing output into a file - could that be the reason? Thanks for your replies. -- Jiøí Voller Laboratory of Growth Regulators Palacký University Institute of Experimental Botany AS CR ©lechtitelù 11, 783 71 Olomouc Czech Republic http://rustreg.upol.cz landline: +420-585-634-855 cell: +420-737-520-506 fax: +420-585-634-870 -- [[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] g(x,y) = f(x,y) - e(x)- e(y)?
Thanks everyone for the help. Greg, I posted the equation from Prof Brillinger's paper. To me it is not clear how or why to subtract a vector from a matrix (since they have diff dimensions). That's why I posted my question to the experts here! Anyway, I have some answers now. Cheers Bill On Wed, Apr 23, 2008 at 5:52 PM, Greg Snow [EMAIL PROTECTED] wrote: It is not clear exactly what you want to do by subtracting a vector from a matrix, but look at the sweep function, it may do what you want. __ 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] sys.time question - how to improve R performance
On Sat, 26 Apr 2008, Jiří Voller wrote: Dear R-users, I run my program a difference between sys.times is as follows: user system elapsed 60167.53 2848.75 63278.93 I am quite puzzled how it may happen that system time is so much shorter than user time, I have 2 core computer most of the time (90%) I was not doing anything else with the computer. What could be the reason of such a difference? I switched off file indexing and all other resident programs.. My program include writing output into a file - could that be the reason? First, do you mean system.time()? Its help page refers you to ?proc.time, which says: An object of class 'proc_time' which is a numeric vector of length 5, containing the user, system, and total elapsed times for the currently running R process, and the cumulative sum of user and system times of any child processes spawned by it on which it has waited. (The 'print' method combines the child times with those of the main process.) and it seems you don't know those terms. At any given time a CPU is running a task for a process, it is either in a system call or in user code (R itself or a package). This is recorded every once in a while (a few ms) and totalled. So your process spent 2848s in system calls. The elapsed time is a little more than the sum of the user and system times, as your R session was not running 100% of the time. You didn't even tell us your OS (see the posting guide) nor your background . On a Unix-alike look at the man pages for time and times. E.g. mine says The tms_utime field contains the CPU time spent executing instructions of the calling process. The tms_stime field contains the CPU time spent in the system while executing tasks on behalf of the calling process. Thanks for your replies. -- Ji?? Voller Laboratory of Growth Regulators Palack? University Institute of Experimental Botany AS CR ?lechtitel? 11, 783 71 Olomouc Czech Republic http://rustreg.upol.cz landline: +420-585-634-855 cell: +420-737-520-506 fax: +420-585-634-870 -- [[alternative HTML version deleted]] -- 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] How to use the optim function if the length of the objective function is greater than 1?
Hi, Please help me out with this.Im a new user of R. Thanks [[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] Ubuntu vs Mac OS X
On 26 April 2008 at 12:27, Prof Brian Ripley wrote: | 3) If you are doing matrix algebra you want an optimized BLAS. The CRAN | build of R for Mac OS by default uses vecLib (which is optimized by | Apple). On other OSes R can be built to use optimized BLAS (see the | manual) -- my understanding is that on Debian derivatives it is built to | make use of the system libblas, and you can install versions of the latter | based on ATLAS. Some optimized BLAS can make use of multiple CPUs. Yes, on Ubuntu, use 'apt-get install atlas3-base' for basic tuned Atlas, or 'apt-get install atlas3-sse2' assuming that sse2 is the closest fit to the cpu in question. 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] quasi-random sequences
You seem to have ambiguous requirements. First, you want equidistribution for a packing structure, which would suggest closest packing or quasirandom sequences, as you have tried. But then you are disturbed by the packing structure, because it gives a perceivable pattern, so you wish to randomize it, but under some unspecified condition of equidistribution (your electrostatic repulsion algorithm). You obviously have some constraints on selection you have not quantified yet. E.g., circles of unspecified radii cannot overlap. You should also realize that any distribution of centers under such constraints will always exhibit structure due to the constraints. Is your problem simply to give an appearance of randomness to the casual observer, or something more definite? You also need to say something about the packing density involved. On the face of it, you with At 06:22 AM 4/26/2008, baptiste Auguié wrote: Dear list useRs, I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, as in: N - 100 positions - matrix(rnorm(2 * N, mean = 0 , sd= 0.5), nrow=N) sizes-rnorm(N, mean = 0 , sd= 1) plot(positions,type=p,cex=sizes) My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, library(gsl) g - qrng_alloc(type=sobol,dim=2) qrng_get(g,n= N) -xy plot((xy),t=p,cex=0.5) but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. Thanks for any comment / suggestion, Baptiste _ Baptiste Auguié Physics Department University of Exeter Stocker Road, Exeter, Devon, EX4 4QL, UK Phone: +44 1392 264187 http://newton.ex.ac.uk/research/emag http://projects.ex.ac.uk/atto __ 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. Robert A. LaBudde, PhD, PAS, Dpl. ACAFS e-mail: [EMAIL PROTECTED] Least Cost Formulations, Ltd.URL: http://lcfltd.com/ 824 Timberlake Drive Tel: 757-467-0954 Virginia Beach, VA 23464-3239Fax: 757-467-2947 Vere scire est per causas scire __ 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 use the optim function if the length of the objectivefunction is greater than 1?
Help you out with what? You have not provided a commented, minimal, self-contained, reproducible example as you were asked to provide in the bottom text of every note to r-help. Can you expect help with a problem only you know? I *suspect* that you will need to define an objective function, such as the sum of the squares of deviations, so that the function is then single-valued. But that's only a guess, based on insufficient information. Charles Annis, P.E. [EMAIL PROTECTED] phone: 561-352-9699 eFax: 614-455-3265 http://www.StatisticalEngineering.com -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of kakul modani Sent: Saturday, April 26, 2008 10:40 AM To: r-help@r-project.org Subject: [R] How to use the optim function if the length of the objectivefunction is greater than 1? Hi, Please help me out with this.Im a new user of R. Thanks [[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.
Re: [R] quasi-random sequences
baptiste Auguié ba208 at exeter.ac.uk writes: Dear list useRs, You might be interested to apply the Hammersley or Halton point sets that are often used in numerical integration or Differential Evolution. These pseudo-random distributions are both uniform and irregular, but have a kind of minimum resolution There is an implementation of Halton Sequences in the often overlooked 'sfsmisc' package, see the 'QUnif()' function there. The help includes a visualization example dispaying the behavior I think you were looking for. Hans Werner I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, [...] My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, [...] but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. Thanks for any comment / suggestion, Baptiste __ 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] matrix from list
Olivier Lefevre wrote: Anyway you are right that it would still return the kind of object, only subsetted, which is not I want. I mean [] would do that; I know [[]] doesn't. Yet I still don't see why one accepts vector arguments but not the other: they are both indexing operators. It is such inconsistencies that make languages hard to learn. -- O.L. __ 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] Consistency of m-shapiro.test
Hello all, I tried several experiments with the mshapiro.test package in R and compared it with the energy package to test for multivariate normality and find that the mshapiro.test is not consistent which is a bit concerning and has suspicious behavior. On the other hand the energy test seems to be a more appropriate test for testing multivariate normality in any dimension. I looked for the reference literature and either seem to reference the original literature on the univariate Shapiro Test or implementation algorithm of the Shapiro test. I was wondering if the procedure used in the mshapiro.test is published elsewhere. For direct comparison we selected three alternative distributions used in Quiroz and Manzotti (2001): 1) Gaussian mixture-bimodal (GM) - (1=2)N (0; Iq) + (1=2)N (3; Iq), 2) Uniformdistribution on the unit cube [0; 1]q, and 3) distribution having i.i.d. coordinates with the Logistic distribution. We evaluate the power of each statistic(1) energy statistic of Szekely and Rizzo, and 2) Multivariate Shapiro Wilk, on the set of alternatives considered, from sets of 1000 samples with the alternative distribution at each dimension q = 2; 3; 4; 5 and each sample size n = 20; 50; 100. We use the R implementation of the energy test and multivariate Shapiro-Wilk test. It seems as if the Multivariate Shapiro-Wilk test exhibits good power against alternatives that are Gaussian Mixture. My numbers do not agree with the power reportedby Szekely and Rizzo in their paper for the Gaussian Mixture distribution (but this is of less importance). I find behavior that are kind of suspect with the Shapiro-Wilk method, such as increase power with smaller sample size for U[0; 1]q and increase power across all alternatives for increasing q. On the other hand, the multivariate energy test exhibits consistent power against all alternatives, i.e. power increases with sample size which is what is desired from any test. Best Regards, Manuel Rivas [[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] matrix from list
Martin Maechler wrote: The difference to as.matrix() is that data.matrix() also produces a numeric matrix in the case the data frame contains factors. Thanks, that is useful but it is becoming a little rococo: so may ways to do this. Also, what if I have a list, not a data frame? read.table returns a data frame but a select from a spreadsheet (via RODBC) returns a list. - I don't understand why you want to throw away the dimnames of the resulting numeric matrix. Rather, in general you'd want to keep them. Merely because I find it confusing. When printing, which you do a lot of when developing in interactive languages like R, list and as.matrix(list) look absolutely the same, which is irksome and confusing since you have to resort to mode to check what you've got, under penalty of strange errors farther down. So I stick to the convention of unattributed matrices and arrays in my code: that way it's easy to know which is which. Regards, -- O.L. __ 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] nested anova and multiple comparisons
Hello R List: My problem is with a nested anova. I have read the r-help and it has answered some of my questions but i still need some help on this one. I have also posted for help on this data set before, so i apologize in advance for any repetition. My design is as follows: response: Quadrat Counts (individuals per quadrat) Explanatory: Region (3 regions) Locations (4 locations nested within each region for a total of 12) Site (4 sites nested within each location within each region for a total of 48) I want to analyse this data set as an observational study where i am interested if 1) There is significant variation at each scale in the study 2) partition the variance components for each scale. 3) Conduct multiple comparisons at the highest level in the study (region) I have managed to accomplish my first two goals by analyzing the data as a 3 level nested Anova with the following code mod1 - aov(Count~Region/Location/Site, data=data) This allows me to get the MS for a Anova table. However R does not compute the correct F-statistics (because it uses the residual MS for the denominator in all calculations which it shouldnt) so i manually computed the F-stats and variance components by hand. From reading the help guide i learned about and tried using the Error(Region/Location/Site) command but then i can only get MS and no F-statistics and still hand to compute the rest by hand. My problem now is that i would liek to use TukeyHSD for multiple comparisons. Howeber since R is unable to compute the correct F statistics in this nested design i also think it is using the wrong MS and df in calculating tukeys q. for example when i use TukeyHSD(mod1, Region) i will get values however i do not think they have been calculated correctly. Furthermore when i use the Error(Region/Location/Site) term i can then no longer use TukeyHSD as i get the error message that there is no applicable use of tukey on this object. i am just wondering if there is any way to use Multiple comparisons with R in a nested design like mine. I have thought about using lme or lmer but if i understand them right with a balanced design i should be able to get the same result using aov Thanks __ 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] quasi-random sequences
Thank you all for the great suggestions and comments. As two of you pointed out, the problem was not well defined (who said a well-posed problem is a problem solved?), and also it seems to be a very wide topic. I've had an interesting reading discussing the similarities between half-toning in black and white printing, and the quasi Monte Carlo integration technique [1]. I'm trying the suggested ways plus a few other things, I'll put together a summary of my discoveries when I've clarified it all. Thanks again, baptiste [1] Halftoning and Quasi-Monte Carlo, K. M. Hanson, Sensitivity Analysis of Model Output, K. M. Hanson and F. M. Hemez, eds., pp. 430-442 (Los Alamos Research Library, 2005) _ Baptiste Auguié Physics Department University of Exeter Stocker Road, Exeter, Devon, EX4 4QL, UK Phone: +44 1392 264187 http://newton.ex.ac.uk/research/emag http://projects.ex.ac.uk/atto __ 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] resolution (dpi) problem
I am using R 2.4.1 with Windows XP. I use the plot command in a fairly simple script and I use the right mouse click on the plot and save as a postscript file. I used the resultant file in a paper which was submitted electronically. However, I get the following response from the journal: Your manuscript has been unsubmitted because you failed to meet the submission guidelines as indicated below: -Your figures must be submitted in TIFF or EPS format according to the following minimum resolutions: 1200 dpi for black and white line art (simple bar graphs, charts, etc.) 300dpi for halftones (black and white photographs) 600dpi for combination halftones (Photographs that also contain line art such as labeling or thin lines) Does anyone know how to produce the correct settings for the journal (i.e. 1200 dpi)? Thankyou Chris W [[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] Ubuntu vs. Windows
On Tue, 22 Apr 2008, Doran, Harold wrote: Dear List: I am very much a unix neophyte, but recently had a Ubuntu box installed in my office. I commonly use Windows XP with 3 GB RAM on my machine and the Ubuntu machine is exactly the same as my windows box (e.g., processor and RAM) as far as I can tell. Now, I recently had to run a very large lmer analysis using my windows machine, but was unable to due to memory limitations, even after increasing all the memory limits in R (which I think is a 2gig max according to the FAQ for windows). So, to make this computationally feasible, I had to sample from my very big data set and then run the analysis. Even still, it would take something on the order of 45 mins to 1 hr to get parameter estimates. (BTW, SAS Proc nlmixed was even worse and kept giving execution errors until the data set was very small and then it ran for a long time) However, I just ran the same analysis on the Ubuntu machine with the full, complete data set, which is very big and lmer gave me back parameter estimates in less than 5 minutes. Because I have so little experience with Ubuntu, I am quite pleased and would like to understand this a bit better. Does this occur because R is a bit friendlier with unix somehow? Or, is this occuring because unix somehow has more efficient methods for memory allocation? On the same hardware the differences between windows and linux performance are generally minor, but there are many things that can cause very poor performance on either platform. I wish I knew enough to even ask the right questions. So, I welcome any enlightenment members may add. I have seen very big differences in performance on computational benchmarks for hardware with similar basic specifications (CPU type and clock, RAM, etc). Often the difference is a symptom of broken hardware or some misconfiguration. Can you see the difference in performance in other applications? Here are some things to consider: 1. anti-virus scanning and other background tasks -- I've seen systems configured to scan gigabyte network drives. Windows task manager and linux top, etc. can give an idea of what is using a lot of CPU, but they are not so helpful if the issue involves I/O bottlenecks. 2. incorrect hardware configuration in the system BIOS. This happens far too often, even with big name vendors. I like to run some benchmarks on every new system to make sure there aren't some basic configuration errors, and to have as a reference if I suspect problems after the systems have been in use. 3. network problems. Where I work, so PC's (both linux and Windows) get the ethernet duplex setting wrong when booted. This can result in poor performance when using networked disks without other symptoms. On windows, the repair network connection button often clears the problem. On linux, ethtool can display and change ethernet settings. 4. all sorts of hardware issues -- sometimes useful data appear in the system logs. Use event viewer on Windows, look at /var/log/messages and /var/log/dmesg on linux. 5. does the slow system exhibit a lot more disk activity? Sometimes this is hard to detect, but most systems do provide some statistics. Try running some I/O intensive benchmark at the same time your R job is running. -- George N. White III [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] Use of survreg.distributions
It would seem to me that your problem is in your data. For survreg, you have to have positive durations, and if your data have y=-5 as seen in Error in survreg(Surv(y, y = -5, type = left) ~ x + : Invalid survival times for this distribution In addition: Warning messages: 1: In log(dlist$dtrans(Y[exactsurv, 1])) : NaNs produced 2: In log(y) : NaNs produced then you have negative durations in the data. I would suggest truncating your data at y0, as that would avoid taking natural logs of negative numbers. Corey Sparks Corey S. Sparks, Ph.D. Assistant Professor Department of Demography and Organization Studies University of Texas San Antonio One UTSA Circle San Antonio, TX 78249 email:[EMAIL PROTECTED] Date: Fri, 25 Apr 2008 10:02:21 -0700 (PDT) From: Abdus Sattar [EMAIL PROTECTED] Subject: Re: [R] Use of survreg.distributions To: Terry Therneau [EMAIL PROTECTED] Cc: r-help@R-project.org Message-ID: [EMAIL PROTECTED] Content-Type: text/plain Hello Dr. Therneau: Thank you for your response. Let me explain to you want I want. My y variable(time) is normal and would like to fit the model on logarithmic transformation of y (log(y)). I tried to run codes according to your suggestion: tfit=survreg(Surv(y, y=-5, type=left)~x + cluster(id), dist=lognormal, data=y.data, scale=0, weights=w) The following error message is giving: Error in survreg(Surv(y, y = -5, type = left) ~ x + : Invalid survival times for this distribution In addition: Warning messages: 1: In log(dlist$dtrans(Y[exactsurv, 1])) : NaNs produced 2: In log(y) : NaNs produced Note, the data file y.data does not contain any missing data! Do you know why it is giving me such an error message please? Thank you again for your helpful comment/suggestion. Best Regards, Abdus Sattar [EMAIL PROTECTED] - Original Message From: Terry Therneau [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: r-help@R-project.org Sent: Friday, April 25, 2008 8:46:45 AM Subject: Re: [R] Use of survreg.distributions --begin included message --- I am using survreg(Surv()) for fitting a Tobit model of left-censored longitudinal data. For logarithmic transformation of y data, I am trying use survreg.distributions in the following way: tfit=survreg(Surv(y, y=-5, type=left)~x + cluster(id), dist=gaussian, data=y.data, scale=0, weights=w) my.gaussian-survreg.distributions$gaussian my.gaussian$name=lognormal my.gaussian$dist-my.gaussian tfit=survreg(Surv(y, y=-5, type=left)~x + cluster(id), dist=my.gaussian, data=y.data, scale=0, weights=w) If I run these codes then I got the following error message, Error in survreg(Surv(y, y = -5, type = left) ~ x + : Invalid distribution object Does anybody can help me in identifying the error(s) in these code please? --- end include Can you tell us what you are trying to do? Your first model was a fit of y ~ x + eps, eps ~ Gaussian. If what you want is log(y) ~ x + eps, then all that you need do is use dist=loggaussian in the survreg call. (Or 'lognormal'; which is the same distribution.) 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] R question about prediction from a fitted model
Lisa [EMAIL PROTECTED] wrote in news:[EMAIL PROTECTED]: Hi Jorge, ... the new question is, is there any prediction for the cox model? I tried predict(Surv(y~x),newdata), it only gave the fitted values. Any clue on this? thanks a lot! I cannot claim great expertise in this area, but my impression is that a Surv() objects is not really a Cox model, but rather an object that can be an argument to one of the Cox PH functions. Suggestion: construct an example and remember to specify which of the libraries the Cox function came from. -- David Winsemius __ 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] nested anova and multiple comparisons
Hi Stephen, On Sat, Apr 26, 2008 at 10:29 AM, Stephen Cole [EMAIL PROTECTED] wrote: ...snip... I have managed to accomplish my first two goals by analyzing the data as a 3 level nested Anova with the following code mod1 - aov(Count~Region/Location/Site, data=data) This allows me to get the MS for a Anova table. However R does not compute the correct F-statistics (because it uses the residual MS for the denominator in all calculations which it shouldnt) so i manually computed the F-stats and variance components by hand. R's just doing what it was asked it to do. The fact that it used the residual MS isn't a surprise given your code. From reading the help guide i learned about and tried using the Error(Region/Location/Site) command but then i can only get MS and no F-statistics and still hand to compute the rest by hand. Yes, if you want to use Method-of-Moments estimation to solve for expected mean squares for the variance components and calculate F-stats w/ the 'correct' MS, then then specifying error strata is reasonable for balanced data (assuming you've met model assumptions). However, I believe most statisticians these days are more inclined to use (Restricted) Maximum Likelihood estimation (e.g. using lme or lmer) because of the added fliexibility it provides (also it won't produce negative variance estimates when the mean squared error is larger than the mean square between the groups of interest) As for why you are only getting MS and no F-stats, it's hard to say without a reproducible example (see the posting guide). In my experience this will occur when there are insufficient degrees of freedom for calculating an F-stat at a given level. My problem now is that i would liek to use TukeyHSD for multiple comparisons. Howeber since R is unable to compute the correct F statistics in this nested design i also think it is using the wrong MS and df in calculating tukeys q. for example when i use Again, it's not that R is unable to, it's that you asked R not to. TukeyHSD(mod1, Region) i will get values however i do not think they have been calculated correctly. Furthermore when i use the Error(Region/Location/Site) term i can then no longer use TukeyHSD as i get the error message that there is no applicable use of tukey on this object. Looking at the methods available for TukeyHSD shows that it will work for an object of class aov methods(TukeyHSD) [1] TukeyHSD.aov But ?aov states: Value: An object of class 'c(aov, lm)' or for multiple responses of class 'c(maov, aov, mlm, lm)' or for multiple error strata of class 'aovlist'. There are 'print' and 'summary' methods available for these. So your model with error strata is of class aovlist not aov. i am just wondering if there is any way to use Multiple comparisons with R in a nested design like mine. I'm not sure but the package 'multcomp' might be of help. I have thought about using lme or lmer but if i understand them right with a balanced design i should be able to get the same result using aov Even with balanced data, you don't get the exact same thing with aov. As mentioned above lme and lmer use different estimation methods. One of the advantages of using (RE)ML is a lot more flexibility in how you specifiy the structure of random effects covariance matrices, and (at least with lme) you have a lot of flexibility in how you structure the residual covariance matrix as well. This is particularly useful when you are not meeting assumptions of constant variance and/or independence of observations (which seems to be the rule rather than the exception with ecological data with a spatial component, such as you appear to have). The theory behind analyses you want to do are quite complex with many potential pitfalls. If at all possible, I highly recommend finding the help of someone who is an expert in these types of models (known as mixed, hierarchical, multilevel, random effects, variance components, nested ANOVA, etc). best, Kingsford Jones Thanks __ 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] resolution (dpi) problem
See comments inline: Chris Walker wrote: I am using R 2.4.1 with Windows XP. First, you are using a version of R that is a year and a half and 6 releases out of date. Version 2.7.0 was just released this past week. You can download it from your nearest CRAN mirror. I use the plot command in a fairly simple script and I use the right mouse click on the plot and save as a postscript file. I used the resultant file in a paper which was submitted electronically. However, I get the following response from the journal: Your manuscript has been unsubmitted because you failed to meet the submission guidelines as indicated below: -Your figures must be submitted in TIFF or EPS format according to the following minimum resolutions: 1200 dpi for black and white line art (simple bar graphs, charts, etc.) 300dpi for halftones (black and white photographs) 600dpi for combination halftones (Photographs that also contain line art such as labeling or thin lines) Does anyone know how to produce the correct settings for the journal (i.e. 1200 dpi)? Their comments about resolution apply to TIFF files and not EPS files, which are resolution independent. It has been several years since I last used the Windows version of R, but if the 'File Save As' menu for the plot indicates Postscript and not Encapsulated Postscript, that is your problem. You can use dev.copy2eps(...) after you have plotted the graphic to the screen device, or better, plot directly to an EPS file by surrounding your plot code with: postscript(FileName.eps, width = X, height = Y, paper = special, horizontal = FALSE, onepage = FALSE) YOUR PLOT CODE HERE dev.off() See ?dev.copy and/or ?postscript for more help. And...be sure to install the latest version of R. :-) 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] resolution (dpi) problem
On 26-Apr-08 19:30:35, Chris Walker wrote: I am using R 2.4.1 with Windows XP. I use the plot command in a fairly simple script and I use the right mouse click on the plot and save as a postscript file. I used the resultant file in a paper which was submitted electronically. However, I get the following response from the journal: Your manuscript has been unsubmitted because you failed to meet the submission guidelines as indicated below: -Your figures must be submitted in TIFF or EPS format according to the following minimum resolutions: 1200 dpi for black and white line art (simple bar graphs, charts, etc.) 300dpi for halftones (black and white photographs) 600dpi for combination halftones (Photographs that also contain line art such as labeling or thin lines) Does anyone know how to produce the correct settings for the journal (i.e.1200 dpi)? Thankyou Chris W I'm about to swim in (for me) murky waters here, since I don't use R on Windows, so things may happen on that platform which I'm not aware of. But I just want to make general comments about PostScript and R. 1. R's postscript() device produces EPS, so that bit should be satisfied. 2. Normally, except when a graphic has been converted from a bit-mapped format, a PostScript (or EPS) file does not have any intrinsic resolution, so long as what it represents is vector graphics (which includes the rendering of letters, numerals, symbols, etc.). Any resolution applying to the result when a PS/EPS file is displayed/printed depends on the resolution of the end device (screen/printer etc.) which does the rendering. In principle, PS/EPS has infinite resolution (for instance, printing from EPS to photograpic film using a laser beam could have resolution as fine as 100,000). 3. It is of course possible that the software generating the graphic implements certain things as bit-maps in the first place, in which case what goes in the PS/EPS file will inevitably do the same. You do not say what sort of graphic you have plotted, so one cannot tell whether (3) applies. However, my feeling is that either the journal has got the wrong impression of what you sent them, or what was intended to be an EPS file in fact got created/converted to some other (bit-mapped) format before you extracted it and sent it off. Sorry not to be more specifically helpful, but I especially wanted to make points (1) and (2) above, for clarification. Best wishes, Ted. E-Mail: (Ted Harding) [EMAIL PROTECTED] Fax-to-email: +44 (0)870 094 0861 Date: 26-Apr-08 Time: 23:22:36 -- 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] quasi-random sequences
You might want to shuffle coordinates independently to get rid of the diagonals. Otherwise what quasi-random sequence guarantee are upper boundaries on the coverage errors, but not anything nice-looking and irregular. Sobol' sequences, even though they are theoretically superior to some others (e.g., Halton sequences more popular among economists), are especially nasty in producing bands and bricks on the low dimensional plots. Among statisticians, Art Owen from Stanford is almost the only one interested in this sort of stuff (referred to as quasi-Monte Carlo, in his field(s)). You might have better luck on a physics list with a question like yours. On Sat, Apr 26, 2008 at 5:22 AM, baptiste Auguié [EMAIL PROTECTED] wrote: Dear list useRs, I have to generate a random set of coordinates (x,y) in [-1 ; 1]^2 for say, N points. At each of these points is drawn a circle (later on, an ellipse) of random size, as in: N - 100 positions - matrix(rnorm(2 * N, mean = 0 , sd= 0.5), nrow=N) sizes-rnorm(N, mean = 0 , sd= 1) plot(positions,type=p,cex=sizes) My problem is to avoid collisions (overlap, really) between the points. I would like some random pattern, but with a minimum exclusion distance. In looking up Numerical recipes in C, I found out about some Sobol quasi-random sequences, which one can call from the gsl package, library(gsl) g - qrng_alloc(type=sobol,dim=2) qrng_get(g,n= N) -xy plot((xy),t=p,cex=0.5) but this does not look very random: I clearly see some pattern (diagonals, etc...), and even the non-overlapping condition is not impressive. One (painful) way I can foresee is to check the distance between each symbol and the others, and move the overlapping ones in a recursive manner. Before delving into this, I wanted to check I'm not overlooking something in the rgl quasi-random sequences, or missing a more obvious way to generate such patterns. Perhaps solving an electrostatic problem with a potential both attractive at long distances and repulsive at short distances is a better way? I have a vague recollection of hearing that somewhere to position points evenly on a sphere. -- Stas Kolenikov, also found at http://stas.kolenikov.name Small print: I don't check Gmail account regularly. __ 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] Calling a stored model within the predict() function
Hi all, First of all, I'm a novice R user (less that a week), so perhaps my code isn't very efficient. Using the MBoost package I created a model using the following command and saved it to a file for later use: model - gamboost(fpfm,data=SampleClusterData,baselearner=bbs) # Creating a model save(model,file=model.RData) # Saving a model After this, during a new R session, I want to deploy this model: setwd(Q:/Program Files/R/library/mboost/data) NewData - read.table(NewData.txt,header=TRUE,sep=,,) #Calling a new data set setwd(Q:/Program Files/R/library/mboost) model - load(model.rdata) # Loading the model This won't work. When I check the content of the 'model' object, the only screen output I see is : model [1]model ...and that the actual file size from 'model.rdata' is exactly 4,100 kb (is this a limit?!) Then I try to call this model object with the predict() function: CRPredict - predict(model, newdata= SampleClusterData) # Make the predictions This results in the following error message: Error in UseMethod(predict) : no applicable method for predict Is the a correct way to store and call a model or is it only possible to create deploy a model during the same R session? Thanks! Jim -- View this message in context: http://www.nabble.com/Calling-a-stored-model-within-the-predict%28%29-function-tp16918111p16918111.html 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.
[R] reading information
teste2
theta_chapeu f_estrela k a0 a1 a2 a3
13.21.2 3 2 1 4 5
p-function(k){
+ i-1
+ for(i in 1:k){
+ p2=teste2[3+i]}
+ return(p2)}
p3-p(teste2$k)
p3
a2
1 4
p3[1]
a2
1 4
how i can make an array with the polinomyal function?
lilke p3[1]=1 p3[2]=4 p3[3]=5
tks
[[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] Lists of data.frame
I cannot find out how to build data structures as lists of data structures. I want to do... r-help@r-project.org d1 - data.frame(x=1, y=2) d2 - data.frame(x=1, y=3) d3 - data.frame(x=21, y=3) q1 - data.frame(q=d1, n=a) q2 - data.frame(q=d2, n=a) q3 - data.frame(q=d3, n=a) v - vector() or list() or whatever v[1] - q1 v[2] - q2 v[3] - q3 then access d2$x as: v[2]$q$x But I cannot get any thing like this. I can find no examples in the manuals. cheers Worik [[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] reading information
p-function(x){
+ for(i in 1:teste2[k]+1){
+ p2=c+teste2[3+i]*x^(i-1)
+ c=p2}
+ return(c)}
p3-p(x)
if i make this i can use a p(x) calling a function in anohter function? no ,
right?
[[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] system.time()
I thought system.time() would return three numbers, the first two adding up to the third one. However, this does not hold in my system running Windows Vista Home Premium, with an Intel Core 2 Duo processor. For example, using the code in the help for system.time() (with one zero added), I got: system.time(for(i in 1:100) mad(runif(1))) user system elapsed 0.270.000.25 Is this a problem with my system clock or am I missing something? Thanks. FS __ 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] Lists of data.frame
I think this is what you want. You need to look at the help file for ?'[' and ?'[['. Also understand the differences between a data.frame and a list. in the case of defining qn you need to use a list and not a data.frame because here is what happens with a data.frame: q1 - data.frame(q=d1, n=a) str(q1) 'data.frame': 1 obs. of 3 variables: $ q.x: num 1 $ q.y: num 2 $ n : Factor w/ 1 level a: 1 Which is probably not want you want. So the following should give you the results you want d1 - data.frame(x=1, y=2) d2 - data.frame(x=1, y=3) d3 - data.frame(x=21, y=3) q1 - list(q=d1, n=a) q2 - list(q=d2, n=a) q3 - list(q=d3, n=a) v - list() v[[1]] - q1 v[[2]] - q2 v[[3]] - q3 str(v) List of 3 $ :List of 2 ..$ q:'data.frame': 1 obs. of 2 variables: .. ..$ x: num 1 .. ..$ y: num 2 ..$ n: chr a $ :List of 2 ..$ q:'data.frame': 1 obs. of 2 variables: .. ..$ x: num 1 .. ..$ y: num 3 ..$ n: chr a $ :List of 2 ..$ q:'data.frame': 1 obs. of 2 variables: .. ..$ x: num 21 .. ..$ y: num 3 ..$ n: chr a #access v[[2]]$q$x [1] 1 On Sat, Apr 26, 2008 at 8:04 PM, Worik R [EMAIL PROTECTED] wrote: I cannot find out how to build data structures as lists of data structures. I want to do... r-help@r-project.org d1 - data.frame(x=1, y=2) d2 - data.frame(x=1, y=3) d3 - data.frame(x=21, y=3) q1 - data.frame(q=d1, n=a) q2 - data.frame(q=d2, n=a) q3 - data.frame(q=d3, n=a) v - vector() or list() or whatever v[1] - q1 v[2] - q2 v[3] - q3 then access d2$x as: v[2]$q$x But I cannot get any thing like this. I can find no examples in the manuals. cheers Worik [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem you are trying to solve? __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] integration error when I use optim and integrate simultaneously
Dear R users,
When I use two functions, 'optim' and 'integrate', simultaneously, I always
get an error like this
--
numint = function(z) {
dlnorm(z,mu[1],sqrt(exp(g[1]))) *
dnorm((z-mu[2])/sqrt(exp(g[2])))/sqrt(exp(g[2]))
}
integrate(numint,lower=0,upper=Inf)$value
Error in integrate(numint, lower = 0, upper = Inf) : non-finite function
value
-
Dr. Ripley said for this problem long time ago
You are trying to use a derivative-based optimization method without
supplying derivatives. This will use numerical approximations to the
derivatives, and your objective function will not be suitable as it is
internally using adaptive numerical quadrature and hence is probably not
close enough to a differentiable function (it may well have steps).
So, I am trying to avoid this 'integrate' in R, but I have no idea.
Any suggestion will be greatly appreciated.
Kathryn Lord
--
View this message in context:
http://www.nabble.com/integration-error-when-I-use-%22optim%22-and-%22integrate%22-simultaneously-tp16918109p16918109.html
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] system.time()
The first two numbers will not add up to the third. The third is the
elapsed time (wall clock) and the first two are CPU time; even though
they have what appears to be the same units (seconds), they can not
really be compared. If, for example, was script was taking input from
the console, you will get something like this:
system.time({
+ x - scan()
+ })
1: 12345
2:
Read 1 item
user system elapsed
0.000.067.34
Here it took 7 seconds for me to enter the data, on only 0.06 seconds
of CPU (all system time -- probably the 'read' function).
The differences in your numbers were the CPU is larger than the
elapsed might be due to the resolution of the time since it was a very
short time, or if you are running an a multiprocessor, and the
application can run multithreaded, you will see the same thing happen
-- e.g., if you have 2 CPU and the application could multithread and
it was CPU bound, then in 10 seconds of elapsed time you might see 20
seconds of CPU time being used.
HTH
On Sat, Apr 26, 2008 at 8:32 PM, Fernando Saldanha [EMAIL PROTECTED] wrote:
I thought system.time() would return three numbers, the first two
adding up to the third one. However, this does not hold in my system
running Windows Vista Home Premium, with an Intel Core 2 Duo
processor. For example, using the code in the help for system.time()
(with one zero added), I got:
system.time(for(i in 1:100) mad(runif(1)))
user system elapsed
0.270.000.25
Is this a problem with my system clock or am I missing something?
Thanks.
FS
__
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.
--
Jim Holtman
Cincinnati, OH
+1 513 646 9390
What is the problem you are trying to solve?
__
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] matrix from list
So you want mylist[[1:2]] to return something (other than an error) when mylist is a list. What if mylist - list( 1:10, 101:110 , some.other.things) so the first 2 elements are vectors of length 10. then mylist[1:2] makes sense as still being a list with the 2 vectors. What should mylist[[1:2]] return in this case? One vector of length 20? or should it return a matrix with 2 columns and 10 rows. Both of those make sense, how should the computer decide between them (it may be obvious to you knowing the context, but how can the computer decide). You can do either of these in R by giving the computer a bit more information (as.matrix or unlist). What if one of the vectors is character and one is numeric, what should the return object be? What if the first element of mylist is the return object from lm and the second element is a function, what should mylist[[1:2]] return then? If you can come up with a set of rules that will cover every possible case, then someone may be willing to implement those rules. But while it is not obvious what to return without giving extra information, it is better to require the extra information through other functions. From: [EMAIL PROTECTED] on behalf of Olivier Lefevre Sent: Sat 4/26/2008 9:43 AM To: [EMAIL PROTECTED] Subject: Re: [R] matrix from list Olivier Lefevre wrote: Anyway you are right that it would still return the kind of object, only subsetted, which is not I want. I mean [] would do that; I know [[]] doesn't. Yet I still don't see why one accepts vector arguments but not the other: they are both indexing operators. It is such inconsistencies that make languages hard to learn. -- O.L. __ 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] integration error when I use optim and integrate simultaneously
On Sat, 26 Apr 2008, kathie wrote:
Dear R users,
When I use two functions, 'optim' and 'integrate', simultaneously, I always
get an error like this
--
numint = function(z) {
dlnorm(z,mu[1],sqrt(exp(g[1]))) *
dnorm((z-mu[2])/sqrt(exp(g[2])))/sqrt(exp(g[2]))
}
integrate(numint,lower=0,upper=Inf)$value
Error in integrate(numint, lower = 0, upper = Inf) : non-finite function
value
-
Dr. Ripley said for this problem long time ago
You are trying to use a derivative-based optimization method without
supplying derivatives. This will use numerical approximations to the
derivatives, and your objective function will not be suitable as it is
internally using adaptive numerical quadrature and hence is probably not
close enough to a differentiable function (it may well have steps).
So, I am trying to avoid this 'integrate' in R, but I have no idea.
Any suggestion will be greatly appreciated.
The implied advice in the quotation above is to use the derivatives.
The function you have is the product of a lognormal and a normal density.
You need the derivatives of those w.r.t their means and variances. You can
probably look these up someplace if the challenge of hacking them out is
too daunting. (Hint: how do you find the MLEs for these densities?)
You can appreciate why integrate() sometimes complains if you spend some
time looking at
curve( numint, 0, uplim )
for suitably chosen values of mu and g and uplim. Say
mu - c(1,1)
g - c(10,1)
uplim - 0.01.
But replacing integrate with a better quadrature rule for this particular
problem probably will not smooth it out enough to substitute for
derivatives.
HTH,
Chuck
Kathryn Lord
--
View this message in context:
http://www.nabble.com/integration-error-when-I-use-%22optim%22-and-%22integrate%22-simultaneously-tp16918109p16918109.html
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.
Charles C. Berry(858) 534-2098
Dept of Family/Preventive Medicine
E mailto:[EMAIL PROTECTED] UC San Diego
http://famprevmed.ucsd.edu/faculty/cberry/ La Jolla, San Diego 92093-0901
__
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.
