[R] rev matrix
Hello I am looking for a function which performs the same thing as rev but on a matrix rather then a vector, based on one column. I'd be happy to read the man page if such a function exists. Thanks, Vera Lindsay __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Gumbel distribution
Does R have built in Gumbel distribution (pdf, ecdf, hazard, parameters estimation) for the minimum case? Thanks Anne [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] rev matrix
On Thu, 14 Oct 2004 [EMAIL PROTECTED] wrote: Hello I am looking for a function which performs the same thing as rev but on a matrix rather then a vector, based on one column. I'd be happy to read the man page if such a function exists. If I understand you, it is easy to do by indexing, but `based on one column' puzzles me and I take it you want each column reversed. A - matrix(1:20, 4, 5) A[4:1, ] or more generally A[rev(seq(len=nrow(A))), ] If that is not what you mean, please give us an example to look at. -- 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 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Gumbel distribution
please, take a look at package evd Matthias Does R have built in Gumbel distribution (pdf, ecdf, hazard, parameters estimation) for the minimum case? Thanks Anne [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] maximum likelihood estimation
Roger == Roger D Peng [EMAIL PROTECTED]
on Thu, 14 Oct 2004 17:06:25 -0400 writes:
Roger What lead you to believe that mle() is defunct? It's
Roger still in the `stats4' package in my installation of
Roger R.
yes indeed. Just to clarify possible confusions:
The 'mle' *package* has been merged into the new 'stats4'
package {for 1.9.0 already}. The mle() function has always been
available and there were even discussions on extension /
generalization of its functioning and the methods working on
mle-class objects (produced by mle(...).
Martin
Roger -roger
Roger T. Murlidharan Nair wrote:
Since mle is defunct is there anyother function I can use
for maximum likelihood estimation ? Thanks ../Murli
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Problem with number characters
Spencer == Spencer Graves [EMAIL PROTECTED] on Thu, 14 Oct 2004 13:41:24 -0700 writes: Spencer It looks like you have several non-printing Spencer characters. nchar will give you the total number Spencer of characters in each character string. Spencer strsplit can break character strings into single Spencer characters, and %in% can be used to classify Spencer them. and you give nice coding examples: Spencer Consider the following: x - Draszt 0%/1ÂÂiso8859-15³ nx - nchar(x) x. - strsplit(x, ) length(x.[[1]]) Spencer [1] 29 namechars - c(letters, LETTERS, as.character(0:9), .) just to be precise: If 'namechars' is supposed to mean ``characters valid in R object names'', then you should have added _ as well: namechars - c(letters, LETTERS, as.character(0:9), ., _) punctuation - c(,, !, +, *, , |) legalchars - c(namechars, punctuation) and 'legalchars' would have to contain quite a bit more I presume, e.g. $, @, (but that wouldn't have been a reason to write this e-mail..) legalx - lapply(x., function(y)(y %in% legalchars)) x.[[1]][!legalx[[1]]] Spencer [1] % /   -  ³ sapply(legalx, sum) Spencer [1] 17 Spencer Will this give you ideas about what to do what you want? Spencer hope this helps. spencer graves (and this too) Martin Maechler, ETH Zurich Spencer Gabor Grothendieck wrote: Assuming that the problem is that your input file has additional embedded characters added by the data base program you could try extracting just the text using the UNIX strings program: strings myfile.csv myfile.txt and see if myfile.txt works with R and if not check out what the differences are between it and the .csv file. Date: Thu, 14 Oct 2004 11:31:33 -0700 From: Scott Waichler [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [R] Problem with number characters I am trying to process text fields scanned in from a csv file that is output from the Windows database program FileMakerPro. The characters onscreen look like regular text, but R does not like their underlying binary form. For example, one of text fields contains a name and a number, but R recognizes the number as something other than what it appears to be in plain text. The character string Draszt 03 after being read into R using scan and = becomes Draszt 03 where the 3 is displayed in my R session as a superscript. Here is the result pasted into this email I'm composing in emacs: Draszt 0%/1ÂÂiso8859-15³ Another clue for the knowledgable: when I try to display the vector element causing trouble, I get CHARSXP: Draszt 0%/1ÂÂiso8859-15³ where again the superscipt part is just 3 in my R session. I'm working in Linux, R version 1.9.1, 2004-06-21. Your help will be much appreciated. Scott Waichler Pacific Northwest National Laboratory [EMAIL PROTECTED] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] C/C++
Hi, Unfortunately I have only some beginners hints to this very interesting topic. Writing R extensions provides examples how to execute r functions from within C code. There is also somewhere a doc how to use R as a shared library (?). But I dont remember in which doc exactly. Furthermore Rserv: rserv (http://www.rosuda.org/Rserve/index.shtml) has a C++ API for calling r functions. Eryk. doktora v wrote: Hey everyone, I have been looking for a while for ways to integrate R's wonderful functions into my C++ software, but I have not found anything concrete. So finally, i post to this list to see if anyouse else knows about this, or has done it!? Is it possible? Are there C++ or C R libraries? Or is it sufficiently easy to build them? your help is much appreciated! thanks doktora __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Dipl. bio-chem. Witold Eryk Wolski MPI-Moleculare Genetic Ihnestrasse 63-73 14195 Berlin tel: 0049-30-83875219 __(_ http://www.molgen.mpg.de/~wolski \__/'v' http://r4proteomics.sourceforge.net||/ \ mail: [EMAIL PROTECTED]^^ m m [EMAIL PROTECTED] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] R and Java
Hi Bobby, You are right - functions is a StringBuffer object. I've compiled without the syntax error. The error message is: Error in dev.print( can only read screen device Do you think this means that RTerm doesn't actually produce a Screen plot when called from Java? i.e., functions.append(plot(something)\n); doesn't produce a screen plot anywhere? Regards, Laura - Laura Jane Dow Energy Scitech Ltd, Scitech House, 1 Barnby Road, Knaphill, Woking, Surrey. GU21 2NL [EMAIL PROTECTED] Phone: +44 (0)1483 79 50 58 Direct Phone: +44 (0)1483 79 78 00 Central Fax: +44 (0)1483 79 76 00 Mobile: +44 (0)7763 10 54 07 http://www.enscitech.com/ http://www.enscitech.com [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] cluster analysis
Dear James, sorry, this is not really an answer. I use cutree to obtain clusters from an hclust object. I do not get from the identify help page that identify should do anything like what you expect it to do... I tried it out and to my surprise it behaved as you said, i.e., it indeed does something at least similar to what you want it to do, and that might be useful also for me. However, I wonder where you got the information that identify could be suitable to obtain the hclust clusters. Puzzled, Christian PS: It seems that each value is typed twice because classi is named, and each value is also a name. Try as.vector(classi). (Perhaps a little useful help in the end?) On Fri, 15 Oct 2004, James Foadi wrote: Hello. I wonder if anyone can help me with this. I'm performing cluster analysis by using hclust in stats package. My data are contained in a data frame with 10 columns, named drops. Firs I create a distance matrix using dist: distanxe - dist(drops) Then I perform cluster analysis via hclust: clusters - hclust(distanze) At this point I want to view the tree plot, and use plot: plot(clusters) Then, once decided which clusters to select, I start identify: classi - identify(clusters) and click on all clusters to be selected; I then finish by right-clicking. My understanding is that classi is now a list containing all individual data, grouped in clusters. In my case classi contained 10 objects, simply named [1], [2], etc. To obtain all individual data belonging to one object I thought that would have sufficed to type for instance: classe_01 - classi[[1]] Unfortunately, rather than obtaining a vector, I obtain a numeric where each value is typed twice. Can anyone explain why, or what I've done wrong? Many thanks, james -- Dr James Foadi Structural Biology Laboratory Department of Chemistry University of York YORK YO10 5YW UK __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html *** Christian Hennig Fachbereich Mathematik-SPST/ZMS, Universitaet Hamburg [EMAIL PROTECTED], http://www.math.uni-hamburg.de/home/hennig/ ### ich empfehle www.boag-online.de __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] cluster analysis
ChrisH == Christian Hennig [EMAIL PROTECTED] on Fri, 15 Oct 2004 11:43:53 +0200 (MEST) writes: ChrisH Dear James, ChrisH sorry, this is not really an answer. nor this. I'm answering Christian... ChrisH I use cutree to obtain clusters from an hclust ChrisH object. I do not get from the identify help page ChrisH that identify should do anything like what you ChrisH expect it to do... I tried it out and to my surprise well, the reason is simple: There's been a nice identify.hclust() method for a long time and this is mentioned (including a link to the page) on the ?hclust page. ChrisH it behaved as you said, i.e., it indeed does ChrisH something at least similar to what you want it to ChrisH do, and that might be useful also for me. However, I ChrisH wonder where you got the information that identify ChrisH could be suitable to obtain the hclust clusters. (see above) --- you see: It *does* pay to read documentation carefully ChrisH Puzzled, ChrisH Christian ChrisH PS: It seems that each value is typed twice because ChrisH classi is named, and each value is also a name. Try ChrisH as.vector(classi). (Perhaps a little useful help in ChrisH the end?) or unname(classi) -- which is slightly more expressive in this case and possibly more desirable in other situations. Martin Maechler, ETH Zurich ChrisH On Fri, 15 Oct 2004, James Foadi wrote: Hello. I wonder if anyone can help me with this. I'm performing cluster analysis by using hclust in stats package. My data are contained in a data frame with 10 columns, named drops. Firs I create a distance matrix using dist: distanxe - dist(drops) Then I perform cluster analysis via hclust: clusters - hclust(distanze) At this point I want to view the tree plot, and use plot: plot(clusters) Then, once decided which clusters to select, I start identify: classi - identify(clusters) and click on all clusters to be selected; I then finish by right-clicking. My understanding is that classi is now a list containing all individual data, grouped in clusters. In my case classi contained 10 objects, simply named [1], [2], etc. To obtain all individual data belonging to one object I thought that would have sufficed to type for instance: classe_01 - classi[[1]] Unfortunately, rather than obtaining a vector, I obtain a numeric where each value is typed twice. Can anyone explain why, or what I've done wrong? Many thanks, james -- Dr James Foadi Structural Biology Laboratory Department of Chemistry University of York YORK YO10 5YW UK __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] cluster analysis
On Friday 15 Oct 2004 10:43 am, you wrote: PS: It seems that each value is typed twice because classi is named, and each value is also a name. Try as.vector(classi). (Perhaps a little useful help in the end?) Indeed. I have tried, for example: as.vector(classi[[1]]) and obtained only one set of values. For some strange reason each object of list classi is a named vector where the name of each component is the component itself. By the way, the cutree function you suggested is even more useful for what I want to do. The info on identify() can easily be obtained using help(hclust); you'll find it at the end of the help page. Many thanks, Christian ! J -- Dr James Foadi Structural Biology Laboratory Department of Chemistry University of York YORK YO10 5YW UK __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Announcing books on R - policy {was Statistical analys...}
Vadim == Vadim Ogranovich [EMAIL PROTECTED] on Thu, 14 Oct 2004 10:50:08 -0700 writes: Vadim I thought that maybe authors of books on R should be Vadim allowed (encouraged ?) to announce Vadim availability/revisions of their books via the Vadim R-packages list? Well, most good *) books on R nowadays have a (usually small) R package going along, for datasets and some own functions. *) One notable current exception: Uwe Ligges book __in German__ http://www.statistik.uni-dortmund.de/~ligges/PmitR/ But maybe he'll change that too. Vadim For example I'd be very interested to have another Vadim look at Dr. Torgo's book when it becomes more Vadim complete and I'd appreciate a revision notice via the list. Hmm, I'm really reluctant to change the 'R-packages' policy, particularly because it won't be necessary in general (see above). Further note that we have the 'R Newsletter' with book reviews, too. Further comments are welcome. Martin Maechler Vadim Thanks, Vadim -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Luis Torgo Sent: Wednesday, October 13, 2004 12:03 PM To: Prof Brian Ripley Cc: Vito Ricci; [EMAIL PROTECTED] Subject: Re: [R] Statistical analysis of a large database On Tue, 2004-10-12 at 08:36, Prof Brian Ripley wrote: Luís Torgo, Data Mining with R. Learning by case studies, Maggio 2003 http://www.liacc.up.pt/~ltorgo/DataMiningWithR/ Please note that that reference is not about large datasets, nor about `data mining' in the generally used sense. It has two studies, one incomplete, on linear regression (with 200 samples) and on time series. I would like to add a few information on these incomplete comments on the book I'm writing. The book is unfinished as mentioned on its Web page. It has currently two reasonably finished chapters: an introduction to R and MySQL and a case study. As mentioned in the book, the first case study is small by data mining standards (200 observations) and has the goal of illustrating techniques that are shared by data mining and other disciplines as well as smoothly introducing the reader to R and its power. It addresses data pre-processing techniques, data visualization, model construction (yes, linear regression but also regression trees), and model evaluation, selection and combination, so I think it is a bit incorrect to say that it is about linear regression that corresponds to 5 of the 50 pages of that chapter. The third (unfinished) chapter (2nd case study) is about financial trading. It includes topics like connections to data bases as well as many other components of a knowledge discovery process. Among those components it includes model construction that involves obviously time series models given the nature of the data. The chapter will include other steps like issues concerning moving from predictions into actions, creation of variables from the original time series, etc.. It is currently being re-written and I expect to upload soon a new revised version of this chapter. The book will include at least two further cases studies that will be larger. Still, I would note that the financial trading case study is potentially very large, as it is a problem where data is constantly growing. The final version of that chapter addresses this issue of having a system that is online in the sense that it is receiving new data in real time (also known as mining data streams in the data mining field). I'm sorry for being so long, but I think it is dangerous to try to resume around 200 pages of an unfinished work in two lines of text. Still, all comments on this on going project are very well welcome and I would like to take this opportunity to thank all people that have been sending me encouraging comments/emails. Luis Torgo -- Luis Torgo FEP/LIACC, University of Porto Phone : (+351) 22 607 88 30 Machine Learning Group Fax : (+351) 22 600 36 54 R. Campo Alegre, 823 email : [EMAIL PROTECTED] 4150 PORTO - PORTUGAL WWW : http://www.liacc.up.pt/~ltorgo __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Vadim __ Vadim [EMAIL PROTECTED] mailing list Vadim https://stat.ethz.ch/mailman/listinfo/r-help Vadim PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Testing for normality of residuals in a regression model
Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] cluster analysis
On Friday 15 Oct 2004 11:02 am, you wrote: or unname(classi) -- which is slightly more expressive in this case and possibly more desirable in other situations. Martin Maechler, ETH Zurich Thanks, Martin. I've tried, like you suggested: un_classi - unname(classi) but nothing changed. By typing, for instance: un_classi[[1]] I still obtained twice the values. But, if I type: un_classe_01 - unname(classi[[1]]) the un_classe_01 is an unnamed vector. Cheers, james -- Dr James Foadi Structural Biology Laboratory Department of Chemistry University of York YORK YO10 5YW UK __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Re: Testing for normality of residuals in a regression model
Dear Federico, see: ? shapiro.test(stats) Shapiro-Wilk Normality Test and ? jarque.bera.test(tseries) Jarque-Bera Test They are the most common tests used for normality testing. Ciao Vito Federico Gherardini wrote on Fri Oct 15 14:44:18 CEST 2004: Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico = Diventare costruttori di soluzioni The business of the statistician is to catalyze the scientific learning process. George E. P. Box Visitate il portale http://www.modugno.it/ e in particolare la sezione su Palese http://www.modugno.it/archivio/cat_palese.shtml __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
What about shapiro.test(resid(fit.object)) Stefano On Fri, Oct 15, 2004 at 02:44:18PM +0200, Federico Gherardini wrote: Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] length with missing values
R-help I have a martix with missing values( in which I want the sample size by column) When I : apply(matrix,2,length) I get the length of the vector regardless of missing values. I can't pass an argument to length in apply. Alternatively I could ifelse ( is.na ( matrix [, columns in matrix ] ) , 0 , 1) Is there any easier way? Thank you __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
Hi Frederico, take also a look at the package nortest: help(package=nortest) Best, Dimitris Dimitris Rizopoulos Ph.D. Student Biostatistical Centre School of Public Health Catholic University of Leuven Address: Kapucijnenvoer 35, Leuven, Belgium Tel: +32/16/396887 Fax: +32/16/337015 Web: http://www.med.kuleuven.ac.be/biostat/ http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm. - Original Message - From: Federico Gherardini [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, October 15, 2004 2:44 PM Subject: [R] Testing for normality of residuals in a regression model Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Filling polygons with points
[EMAIL PROTECTED] wrote : Are there any possibilities to fill a polygon with a point pattern or with a symbol pattern like '+' oder '-' instead of shading lines? In `spatstat' there is code for testing whether points belong to a spatial region. The region can consist of any number of polygons, and the polygons may have holes, etc. While splancs has similar functionality, spatstat additionally allows regions defined by a binary image (e.g. a digital approximation to a circle). The code is available both as a function inside.owin(x,y,win) and as a method for [ for point pattern objects. Example of [: library(spatstat) data(demopat) w - demopat$window # an irregular polygonal window plot(w) g - as.mask(w, dimyx=c(20,20)) # make a 20 x 20 grid # enclosing the window g - as.mask(w, eps = 200) # or set absolute grid spacing 200 p - ppp(raster.x(g), raster.y(g), window=w) # convert to point pattern object p - p[, w] # clip points to window plot(p) plot(p, pch=+) # etc To create a window object use owin() or as.owin(). Adrian Baddeley __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] length with missing values
Hi Luis, just use: apply( mat, 2, function(x) length(x[!is.na(x)]) ) I hope it helps. Best, Dimitris Dimitris Rizopoulos Ph.D. Student Biostatistical Centre School of Public Health Catholic University of Leuven Address: Kapucijnenvoer 35, Leuven, Belgium Tel: +32/16/396887 Fax: +32/16/337015 Web: http://www.med.kuleuven.ac.be/biostat/ http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm - Original Message - From: Luis Rideau Cruz [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, October 15, 2004 1:01 PM Subject: [R] length with missing values R-help I have a martix with missing values( in which I want the sample size by column) When I : apply(matrix,2,length) I get the length of the vector regardless of missing values. I can't pass an argument to length in apply. Alternatively I could ifelse ( is.na ( matrix [, columns in matrix ] ) , 0 , 1) Is there any easier way? Thank you __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] length with missing values
Hi Luis, Are you looking for the number of missings in each column? If so, try this:- bob-matrix(c(2:4, NA, 5:7, NA), ncol=2) colSums(is.na(bob)) [1] 1 1 Hope this helps, Ian. mangosolutions R Consulting and Training Services Tel +44 118 902 6620 Fax +44 118 902 6401 -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Luis Rideau Cruz Sent: 15 October 2004 12:01 To: [EMAIL PROTECTED] Subject: [R] length with missing values R-help I have a martix with missing values( in which I want the sample size by column) When I : apply(matrix,2,length) I get the length of the vector regardless of missing values. I can't pass an argument to length in apply. Alternatively I could ifelse ( is.na ( matrix [, columns in matrix ] ) , 0 , 1) Is there any easier way? Thank you __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] length with missing values
Dimitris Rizopoulos [EMAIL PROTECTED] writes: Hi Luis, just use: apply( mat, 2, function(x) length(x[!is.na(x)]) ) or: apply(!is.na(mat),2,sum) -- O__ Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] length with missing values
or even: colSums(!is.na(mat)) Best, Dimitris - Original Message - From: Peter Dalgaard [EMAIL PROTECTED] To: Dimitris Rizopoulos [EMAIL PROTECTED] Cc: Luis Rideau Cruz [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Friday, October 15, 2004 1:43 PM Subject: Re: [R] length with missing values Dimitris Rizopoulos [EMAIL PROTECTED] writes: Hi Luis, just use: apply( mat, 2, function(x) length(x[!is.na(x)]) ) or: apply(!is.na(mat),2,sum) -- O__ Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] length with missing values
On 15 Oct 2004 at 12:01, Luis Rideau Cruz wrote: R-help I have a martix with missing values( in which I want the sample size by column) When I : apply(matrix,2,length) Hi Try apply(matrix, 2, function(x) sum(!is.na(x))) matrix is reserved word, so it is preferable to use other name for your object. Cheers Petr I get the length of the vector regardless of missing values. I can't pass an argument to length in apply. Alternatively I could ifelse ( is.na ( matrix [, columns in matrix ] ) , 0 , 1) Is there any easier way? Thank you __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Petr Pikal [EMAIL PROTECTED] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] length with missing values
On Fri, 15 Oct 2004, Sundar Dorai-Raj wrote: Luis Rideau Cruz wrote: I have a martix with missing values( in which I want the sample size by column) When I : apply(matrix,2,length) I get the length of the vector regardless of missing values. I can't pass an argument to length in apply. Alternatively I could ifelse ( is.na ( matrix [, columns in matrix ] ) , 0 , 1) Is there any easier way? I think you almost have it: colSums(ifelse(is.na(x), 0, 1)) will return the number of non-NA elements in each column of x. colSums(!is.na(x)) is the same, and more obvious, I think. Remember logical values are coerced to 0/1 in an arithmetical context. -- 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 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] R plot problems
At 10:51 2004-10-15 +0100, you wrote: If you are using base grafics, you use the argument srt (see help(par)), How do you get a, say, 45 degree rotation of the axis labels using `srt'? The help page for par's `las' argument says Note that other string/character rotation (via argument srt to par) does not affect the axis labels Henric __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Dear Federico, A problem with applying a standard test of normality to LS residuals is that the residuals are correlated and heterskedastic even if the standard assumptions of the model hold. In a large sample, this is unlikely to be problematic (unless there's an unusual data configuration), but in a small sample the effect could be nontrivial. One approach is to use BLUS residuals, which transform the LS residuals to a smaller set of uncorrelated, homoskedastic residuals (assuming the correctness of the model). A search of R resources didn't turn up anything for BLUS, but they shouldn't be hard to compute. This is a standard topic covered in many econometrics texts. You might consider the alternative of generating a bootstrapped confidence envelope for the QQ plot; the qq.plot() function in the car package will do this for a linear model. I hope this helps, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Federico Gherardini Sent: Friday, October 15, 2004 7:44 AM To: [EMAIL PROTECTED] Subject: [R] Testing for normality of residuals in a regression model Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] R plot problems
On Fri, 2004-10-15 at 07:18, Henric Nilsson wrote: At 10:51 2004-10-15 +0100, you wrote: If you are using base grafics, you use the argument srt (see help(par)), How do you get a, say, 45 degree rotation of the axis labels using `srt'? The help page for par's `las' argument says Note that other string/character rotation (via argument srt to par) does not affect the axis labels Henric See the new R FAQ: 7.27 How can I create rotated axis labels? HTH, Marc Schwartz __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] cluster analysis
James == James Foadi [EMAIL PROTECTED] on Fri, 15 Oct 2004 11:36:14 +0100 writes: James On Friday 15 Oct 2004 11:02 am, you wrote: or unname(classi) -- which is slightly more expressive in this case and possibly more desirable in other situations. Martin Maechler, ETH Zurich James Thanks, Martin. James I've tried, like you suggested: James un_classi - unname(classi) James but nothing changed. By typing, for instance: James un_classi[[1]] of course -- I just chimed in with Christian who proposed as.vector(.) Since your 'classi' is a list with named vector as components, you'd need something like un_classi - lapply(classi, unname) I'm sorry to have added more confusion. OTOH, really, I think you should learn a bit more about basic manipulation of R objects and study something like An Introduction to R. Regards, Martin James I still obtained twice the values. But, if I type: James un_classe_01 - unname(classi[[1]]) James the un_classe_01 is an unnamed vector. (exactly, since it works on the *component* of a list) James Cheers, James james __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] R plot problems
Use text(), for which srt will work. As an example: plot(rnorm(20), rnorm(20), xaxt=n) text(-2:2, rep(par(usr)[3], 5), c(negative two, negative one, zero, one, two), srt=45, xpd=NA, adj=1) HTH, Andy From: Henric Nilsson At 10:51 2004-10-15 +0100, you wrote: If you are using base grafics, you use the argument srt (see help(par)), How do you get a, say, 45 degree rotation of the axis labels using `srt'? The help page for par's `las' argument says Note that other string/character rotation (via argument srt to par) does not affect the axis labels Henric __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] C/C++
Have you checked Section 5.14 (page 64) of R-extension titled Using these functions in your own C code. Section 5.7 tells you about some of the available subroutines. On Fri, 2004-10-15 at 01:31, doktora v wrote: Hey everyone, I have been looking for a while for ways to integrate R's wonderful functions into my C++ software, but I have not found anything concrete. So finally, i post to this list to see if anyouse else knows about this, or has done it!? Is it possible? Are there C++ or C R libraries? Or is it sufficiently easy to build them? your help is much appreciated! thanks doktora __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] edit plots from the ADE4 package
Dear R-Help
i have a newbie question, how edit size fonts and colors in the plots (scatter.acm or
dudi.acm function) of multiple correspondence analysis in the ADE4 Package?
thanks?
Rafael Gutierrez
Estadístico
Unidad de Tecnología Cerro Matoso S.A.
Tel. 4-7723350 Fax. 4-7723236
*
Este correo y sus anexos pueden ser confidenciales y estar protegidos por derechos de
autor.
Estan dirigidos unica y exclusivamente para uso de el (los) destinatario(s).
Si Usted por error lo ha recibido por favor notifiquelo inmediatamente al remitente y
destruyalo de su sistema.
No debe copiar, ni imprimir, ni distribuir este correo o sus anexos, ni usarlos para
proposito alguno ni dar a conocer su contenido a persona alguna.
Las opiniones, conclusiones y otra informacion contenida en este correo no
relacionadas con el negocio oficial de Cerro Matoso, deben entenderse como personales
y de ninguna manera son avaladas por la Compañia.
*
This message and any attachments may be confidential and may...{{dropped}}
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] constrained splines in GLM
Hi, I would like to use constrained splines in a GLM model (Poisson link) to take into account the nonlinear effect of some covariate. The constraints I need are described below. I have several variables that I need concurrently in the same model. I looked at package mgcv but I do not know if/how I can use it in GLM (not GAM) : I could not manage to adapt the mono.con(mgcv) example to GLM. The help for package fda is not complete. Not sure that backSpline(splines) does what I need. isoreg (modreg) seems to do univariate regressions. Some of my covariates are linear. Three covariates (x1,x2 and x3) must be transformed in a decreasing and convex way like this: |o |o | o | o |o |o | o |- Currently, I use exp(-x1/alpha1)+exp(-x2/alpha2)+exp(-x3/alpha3), I try several alpha's and choose the best according to log-likelihood. One variable should have only one local maximum (that is, the derivative should be zero only once, which is at the top), like this: | | TOP | oo | o o |o o |o o o | o o | with bs() or ns() and no constraint, I get: | | TOP | oo |o o o | o o o |o | o o | which is nonsense (note there are very few observations on the left part) I also tried some parametric forms, choosing via log-likelihood. But with four covariates, it is a lot of parameters to try (several hours with little flexible functions). I am looking for something similar to ns or bs (package splines), which are very convenient to place in the formula of a GLM model. I tried them, adjusting knots, but could not manage what I want. Constraints on some derivatives may do the trick, but I do not know how to implement them in R. Any help or comment would be greatly appreciated ! Mayeul KAUFFMANN Université Pierre Mendès France - Grenoble France __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
Thank you very much for your suggestions! The residuals come from a gls model, because I had to correct for heteroscedasticity using a weighted regression... can I simply apply one of these tests (like shapiro.test) to the standardized residuals from my gls model? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] pdf device --- axes labels text disappeared?
Dear R Wizards: Running R 1.9.1. on amd64. Promise- c(0,20,40); Expect- c(0, 20, 0.2*20+.8*40 ); # this omits printing numbers on the axes labels. pdf(file = bug.pdf ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); # this works postscript(file = bug.eps ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); apologies if this has already been noted elsewhere---I looked but did not find it. (probably googled for the wrong terms, if so.) is this my bug, or R's bug or my R installation bug? help appreciated. /iaw __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
John Fox wrote:
Dear Federico,
A problem with applying a standard test of normality to LS residuals is that
the residuals are correlated and heterskedastic even if the standard
assumptions of the model hold. In a large sample, this is unlikely to be
problematic (unless there's an unusual data configuration), but in a small
sample the effect could be nontrivial.
One approach is to use BLUS residuals, which transform the LS residuals to a
smaller set of uncorrelated, homoskedastic residuals (assuming the
correctness of the model).
I'm not sure if this are BLUE residuals, but the following function
transform to a
smaller set of independent, homoscedastic residuals and the calls
shapiro.test:
I've proposed to make this a method for shapiro.test for lm objects,
but it is
not accepted.
shapiro.test.lm
function (obj)
{
eff - effects(obj)
rank - obj$rank
df.r - obj$df.residual
if (df.r 3)
stop(To few degrees of freedom for residual for the test.)
data.name - deparse(substitute(obj))
x - eff[-(1:rank)]
res - shapiro.test(x)
res$data.name - data.name
res$method - paste(res$method, for residuals of linear model)
res
}
Kjetil
A search of R resources didn't turn up anything
for BLUS, but they shouldn't be hard to compute. This is a standard topic
covered in many econometrics texts.
You might consider the alternative of generating a bootstrapped confidence
envelope for the QQ plot; the qq.plot() function in the car package will do
this for a linear model.
I hope this helps,
John
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Federico Gherardini
Sent: Friday, October 15, 2004 7:44 AM
To: [EMAIL PROTECTED]
Subject: [R] Testing for normality of residuals in a regression model
Hi all,
Is it possible to have a test value for assessing the
normality of residuals from a linear regression model,
instead of simply relying on qqplots?
I've tried to use fitdistr to try and fit the residuals with
a normal distribution, but fitdsitr only returns the
parameters of the distribution and the standard errors, not
the p-value. Am I missing something?
Cheers,
Federico
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
--
Kjetil Halvorsen.
Peace is the most effective weapon of mass construction.
-- Mahdi Elmandjra
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
John Fox wrote: Dear Federico, A problem with applying a standard test of normality to LS residuals is that the residuals are correlated and heterskedastic even if the standard assumptions of the model hold. In a large sample, this is unlikely to be problematic (unless there's an unusual data configuration), but in a small sample the effect could be nontrivial. One approach is to use BLUS residuals, which transform the LS residuals to a smaller set of uncorrelated, homoskedastic residuals (assuming the correctness of the model). A search of R resources didn't turn up anything for BLUS, but they shouldn't be hard to compute. This is a standard topic covered in many econometrics texts. You might consider the alternative of generating a bootstrapped confidence envelope for the QQ plot; the qq.plot() function in the car package will do this for a linear model. I hope this helps, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Federico Gherardini Sent: Friday, October 15, 2004 7:44 AM To: [EMAIL PROTECTED] Subject: [R] Testing for normality of residuals in a regression model Hi all, Is it possible to have a test value for assessing the normality of residuals from a linear regression model, instead of simply relying on qqplots? I've tried to use fitdistr to try and fit the residuals with a normal distribution, but fitdsitr only returns the parameters of the distribution and the standard errors, not the p-value. Am I missing something? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Kjetil Halvorsen. Peace is the most effective weapon of mass construction. -- Mahdi Elmandjra __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Problems saving a graphic
Hi! I'm with a little problem when saving a graphic into pdf. The graphic has the label: Milhões de toneladas. The problem is with the accent. After I save the graphic into pdf the õ doesn't appear. Somebody could help me? Thanks, Talita Perciano Costa Leite Graduanda em Ciência da Computação Universidade Federal de Alagoas - UFAL Departamento de Tecnologia da Informação - TCI Construção de Conhecimento por Agrupamento de Dados - CoCADa __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] edit plots from the ADE4 package
Hi
Try to change the parameters in the s.class function corresponding to col (which is 1
by default) and using the parameter font as a small positive integer determining a
text font for characters and hence an interline spacing.
Leonardo Trujillo
Southamtpton Statistical Sciences Research Institute
[EMAIL PROTECTED]
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Rafael Gutierrez
Sent: 15 October 2004 14:54
To: [EMAIL PROTECTED]
Subject: [R] edit plots from the ADE4 package
Dear R-Help
i have a newbie question, how edit size fonts and colors in the plots (scatter.acm or
dudi.acm function) of multiple correspondence analysis in the ADE4 Package?
thanks?
Rafael Gutierrez
Estadístico
Unidad de Tecnología Cerro Matoso S.A.
Tel. 4-7723350 Fax. 4-7723236
*
Este correo y sus anexos pueden ser confidenciales y estar protegidos por derechos de
autor.
Estan dirigidos unica y exclusivamente para uso de el (los) destinatario(s).
Si Usted por error lo ha recibido por favor notifiquelo inmediatamente al remitente y
destruyalo de su sistema.
No debe copiar, ni imprimir, ni distribuir este correo o sus anexos, ni usarlos para
proposito alguno ni dar a conocer su contenido a persona alguna.
Las opiniones, conclusiones y otra informacion contenida en este correo no
relacionadas con el negocio oficial de Cerro Matoso, deben entenderse como personales
y de ninguna manera son avaladas por la Compañia.
*
This message and any attachments may be confidential and may...{{dropped}}
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] does rpart has similar function as edit.tree?
Hi, list, The 'tree' package has edit.tree to change split variables by user. Does 'rpart' package have something similar? I couldn't find either in the manual or archive. Thanks in advance. Best, Auston __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] pdf device --- axes labels text disappeared?
Dear R Wizards: Running R 1.9.1. on amd64.
Promise- c(0,20,40); Expect- c(0, 20, 0.2*20+.8*40 );
# this omits printing numbers on the axes labels.
pdf(file = bug.pdf );
plot(Promise, Expect, type=b, ylim=c(0,60));
dev.off();
# this works
postscript(file = bug.eps );
plot(Promise, Expect, type=b, ylim=c(0,60));
dev.off();
Dear Ivo,
although I do not want to consider myself as an R wizard, but rather a
helpful soul and because you have provided R-code suited for direct
execution as it should be :-), here is my answer:
in both devices the numbers are printed below the axes. This is on:
platform i386-pc-mingw32
arch i386
os mingw32
system i386, mingw32
status
major2
minor0.0
year 2004
month10
day 04
language R
HTH,
Bernhard
ps: Could the problem be related to your pdf encodings?
apologies if this has already been noted elsewhere---I looked but did
not find it. (probably googled for the wrong terms, if so.) is this
my bug, or R's bug or my R installation bug? help appreciated.
/iaw
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
The information contained herein is confidential and is inte...{{dropped}}
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] pdf device --- axes labels text disappeared?
On Fri, 2004-10-15 at 09:13, [EMAIL PROTECTED] wrote: Dear R Wizards: Running R 1.9.1. on amd64. Promise- c(0,20,40); Expect- c(0, 20, 0.2*20+.8*40 ); # this omits printing numbers on the axes labels. pdf(file = bug.pdf ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); # this works postscript(file = bug.eps ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); apologies if this has already been noted elsewhere---I looked but did not find it. (probably googled for the wrong terms, if so.) is this my bug, or R's bug or my R installation bug? help appreciated. /iaw Hey Ivo! No problems here using R 2.0.0 on 32 bit Intel FC2. Both plots are fine. I don't have a 1.9.1 install at the moment, but I would guess that we would have heard about the problem prior to this. That suggests a possible issue with your install or a change of some sort with 64 bit R. Though again, I suspect other 64 bit users would have commented on this already... HTH, Marc Schwartz __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Problem with number characters
Note that there are also regexp classes that define certain character sets, most notably [:graph:] , which can make it easy to create appropriate regexps. More is in ?regex . Martin Maechler maechler at stat.math.ethz.ch writes: : : Spencer == Spencer Graves spencer.graves at pdf.com : on Thu, 14 Oct 2004 13:41:24 -0700 writes: : : Spencer It looks like you have several non-printing : Spencer characters. nchar will give you the total number : Spencer of characters in each character string. : : Spencer strsplit can break character strings into single : Spencer characters, and %in% can be used to classify : Spencer them. : : and you give nice coding examples: : : Spencer Consider the following: : x - Draszt 0%/1iso8859-15 : nx - nchar(x) : x. - strsplit(x, ) : length(x.[[1]]) : Spencer [1] 29 : : namechars - c(letters, LETTERS, as.character(0:9), .) : : just to be precise: If 'namechars' is supposed to mean : ``characters valid in R object names'', then you should have : added _ as well: : : namechars - c(letters, LETTERS, as.character(0:9), ., _) : : punctuation - c(,, !, +, *, , |) : legalchars - c(namechars, punctuation) : : and 'legalchars' would have to contain quite a bit more I : presume, e.g. $, at , : (but that wouldn't have been a reason to write this e-mail..) : : legalx - lapply(x., function(y)(y %in% legalchars)) : x.[[1]][!legalx[[1]]] : Spencer [1] % / - : : sapply(legalx, sum) : Spencer [1] 17 : : Spencer Will this give you ideas about what to do what you want? : Spencer hope this helps. spencer graves : : (and this too) : : Martin Maechler, ETH Zurich : : : Spencer Gabor Grothendieck wrote: : : Assuming that the problem is that your input file has : additional embedded characters added by the data base : program you could try extracting just the text using : the UNIX strings program: : : strings myfile.csv myfile.txt : : and see if myfile.txt works with R and if not check out : what the differences are between it and the .csv file. : : Date: Thu, 14 Oct 2004 11:31:33 -0700 : From: Scott Waichler scott.waichler at pnl.gov : To: r-help at stat.math.ethz.ch : Subject: [R] Problem with number characters : : : I am trying to process text fields scanned in from a csv file that is : output from the Windows database program FileMakerPro. The characters : onscreen look like regular text, but R does not like their underlying binary form. : For example, one of text fields contains a name and a number, but : R recognizes the number as something other than what it appears : to be in plain text. The character string Draszt 03 after being : read into R using scan and = becomes Draszt 03 where the 3 is : displayed in my R session as a superscript. Here is the result pasted : into this email I'm composing in emacs: Draszt 0%/1iso8859- 15 : Another clue for the knowledgable: when I try to display the vector element : causing trouble, I get : CHARSXP: Draszt 0%/1iso8859-15 : where again the superscipt part is just 3 in my R session. I'm working in : Linux, R version 1.9.1, 2004-06-21. Your help will be much appreciated. : : Scott Waichler : Pacific Northwest National Laboratory : scott.waichler at pnl.gov : : __ : R-help at stat.math.ethz.ch mailing list : https://stat.ethz.ch/mailman/listinfo/r-help : PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html : : __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Within-Subjects ANOVA TukeyHSD
Is there a problem with using TukeyHSD for individual comparisons of means coming from a within-subjects ANOVA? I always get: no applicable method for TukeyHSD Thank you, Steffen __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] pdf device --- axes labels text disappeared?
No problems under R 2.0.0 or 1.9.1 on that platform. Are you *sure* this is a bug in the PDF and not in your pdf reader? Please read the PDF source file to find out. Mine has lines like /F2 1 Tf 12.00 0.00 -0.00 12.00 68.40 47.52 Tm (0) Tj /F2 1 Tf 12.00 0.00 -0.00 12.00 144.39 47.52 Tm (10) Tj /F2 1 Tf 12.00 0.00 -0.00 12.00 223.73 47.52 Tm (20) Tj /F2 1 Tf 12.00 0.00 -0.00 12.00 303.06 47.52 Tm (30) Tj /F2 1 Tf 12.00 0.00 -0.00 12.00 382.39 47.52 Tm (40) Tj On Fri, 15 Oct 2004 [EMAIL PROTECTED] wrote: Dear R Wizards: Running R 1.9.1. on amd64. Promise- c(0,20,40); Expect- c(0, 20, 0.2*20+.8*40 ); # this omits printing numbers on the axes labels. pdf(file = bug.pdf ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); # this works postscript(file = bug.eps ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); apologies if this has already been noted elsewhere---I looked but did not find it. (probably googled for the wrong terms, if so.) is this my bug, or R's bug or my R installation bug? help appreciated. /iaw __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- 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 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] pdf device --- axes labels text disappeared?
/iaw, * [EMAIL PROTECTED] [2004-Oct-15 06:13 AKDT]: Dear R Wizards: Running R 1.9.1. on amd64. Promise- c(0,20,40); Expect- c(0, 20, 0.2*20+.8*40 ); # this omits printing numbers on the axes labels. pdf(file = bug.pdf ); plot(Promise, Expect, type=b, ylim=c(0,60)); dev.off(); Version 1.9.1 (2004-06-21), AMD64, Debian amd64 unstable port. The plot I produced using these commands had all the numbers and lables. I used 'xpdf' version 3.00 to display the plot. Maybe it's your PDF viewer, rather than R? Chris -- Christopher S. Swingley email: [EMAIL PROTECTED] (work) Intl. Arctic Research Center[EMAIL PROTECTED] (personal) University of Alaska Fairbanks www.frontier.iarc.uaf.edu/~cswingle/ __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] one more Rcmdr problem
I did experience the same problem. After installing R 2.0.0 patched and downloading the source for Rcmdr_0.99-12 from John Fox's Web page http://socserv.socsci.mcmaster.ca/jfox/Misc/Rcmdr/ and recompiling the package it now works. Ctrl-C, Ctrl-X, and Ctrl-V, in the text boxes of Rcmdr now work. Adrian Dusa wrote: Hello, I'm using R 2.0.0 with the latest Rcmdr package installed from CRAN, on Windows XP Professional. When trying to copy some commands or results, either from the upper or lower text window, this causes Rcmdr to crash: R for Windows GUI front-end has encountered a problem and needs to close Did anyone have the same problem? I don't think it's my system, as it happened to reinstall my Windows just a few days ago, and the same problem occurred in the former one. Regards, Adrian Adrian Dusa Romanian Social Data Archive 1 Schitu Magureanu Bd. 050025 Bucharest sector 5 Tel. +40 21 3126618\ +40 21 3153122/ int.101 -- Erich Neuwirth, Computer Supported Didactics Working Group Visit our SunSITE at http://sunsite.univie.ac.at Phone: +43-1-4277-38624 Fax: +43-1-4277-9386 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
Berton Gunter wrote: Quite right, John! I have 2 additional questions: 1) Why test for normality of residuals? Suppose you reject -- then what? (residual plots may give information on skewness, multi-modality, data anomalies that can affect the data analysis). Because I want to know if my model satisfies the basic assumptions of regression theory... in other words I want to know if I can trust my model. Cheers, Federico 2) Why test for normality? Is it EVER useful? Suppose you reject -- then what? (I am tempted to add a 3rd question -- why test at all? -- but that is perhaps too iconoclastic and certainly off topic. Let the hounds remain leashed for now.) Cheers, -- Bert Gunter Genentech Non-Clinical Statistics South San Francisco, CA __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] pdf device --- axes labels text disappeared?
Hi: Thank you everyone. You were all correct. This turns out to be a viewer bug---and even more likely a viewer installation bug, not an xpdf bug, much less an R bug. Apologies for having wasted everyone's time. Regards, /ivo --- ivo welch __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Problem with number characters
Scott, Has anyone suggested yet that some options might need adjusting in the Windows program that wrote the file? In my experience, CSV files from Windows applications are typically pure ASCII (though yours clearly isn't). Another possibility is that the program can export in plain ASCII with the fields in set positions. Sometimes that can be done by capturing output intended for basic printers. Either of those approaches might be simpler than trying to fiddle with the quirky output provided. Reverse engineering is error prone. MHP At 10/15/2004 10:47 AM Friday, you wrote: Note that there are also regexp classes that define certain character sets, most notably [:graph:] , which can make it easy to create appropriate regexps. More is in ?regex . Martin Maechler maechler at stat.math.ethz.ch writes: : : Spencer == Spencer Graves spencer.graves at pdf.com : on Thu, 14 Oct 2004 13:41:24 -0700 writes: : : Spencer It looks like you have several non-printing : Spencer characters. nchar will give you the total number : Spencer of characters in each character string. : : Spencer strsplit can break character strings into single : Spencer characters, and %in% can be used to classify : Spencer them. : : and you give nice coding examples: : : Spencer Consider the following: : x - Draszt 0%/1ÂÂiso8859-15³ : nx - nchar(x) : x. - strsplit(x, ) : length(x.[[1]]) : Spencer [1] 29 : : namechars - c(letters, LETTERS, as.character(0:9), .) : : just to be precise: If 'namechars' is supposed to mean : ``characters valid in R object names'', then you should have : added _ as well: : : namechars - c(letters, LETTERS, as.character(0:9), ., _) : : punctuation - c(,, !, +, *, , |) : legalchars - c(namechars, punctuation) : : and 'legalchars' would have to contain quite a bit more I : presume, e.g. $, at , : (but that wouldn't have been a reason to write this e-mail..) : : legalx - lapply(x., function(y)(y %in% legalchars)) : x.[[1]][!legalx[[1]]] : Spencer [1]% /   -  ³ : : sapply(legalx, sum) : Spencer [1] 17 : : Spencer Will this give you ideas about what to do what you want? : Spencer hope this helps. spencer graves : : (and this too) : : Martin Maechler, ETH Zurich : : : Spencer Gabor Grothendieck wrote: : : Assuming that the problem is that your input file has : additional embedded characters added by the data base : program you could try extracting just the text using : the UNIX strings program: : : strings myfile.csv myfile.txt : : and see if myfile.txt works with R and if not check out : what the differences are between it and the .csv file. : : Date: Thu, 14 Oct 2004 11:31:33 -0700 : From: Scott Waichler scott.waichler at pnl.gov : To: r-help at stat.math.ethz.ch : Subject: [R] Problem with number characters : : : I am trying to process text fields scanned in from a csv file that is : output from the Windows database program FileMakerPro. The characters : onscreen look like regular text, but R does not like their underlying binary form. : For example, one of text fields contains a name and a number, but : R recognizes the number as something other than what it appears : to be in plain text. The character string Draszt 03 after being : read into R using scan and = becomes Draszt 03 where the 3 is : displayed in my R session as a superscript. Here is the result pasted : into this email I'm composing in emacs: Draszt 0%/1ÂÂiso8859- 15³ : Another clue for the knowledgable: when I try to display the vector element : causing trouble, I get : CHARSXP: Draszt 0%/1ÂÂiso8859-15³ : where again the superscipt part is just 3 in my R session. I'm working in : Linux, R version 1.9.1, 2004-06-21. Your help will be much appreciated. : : Scott Waichler : Pacific Northwest National Laboratory : scott.waichler at pnl.gov : : __ : R-help at stat.math.ethz.ch mailing list : https://stat.ethz.ch/mailman/listinfo/r-help : PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html : : __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Michael Prager NOAA Center for Coastal Fisheries and Habitat Research Beaufort, North Carolina 28516 USA http://shrimp.ccfhrb.noaa.gov/~mprager/ NOTE: Opinions expressed are personal, not official. No government endorsement of any product is made or implied. __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] add-on package/windows installation probLem
Hi, I have tried to download the bootstrap.tar.gz from the website below: http://cran.r-project.org/src/contrib/Old/0.50/INDEX.html I have extracted the folder to a zip file...and then use the load packages from local zip files. I use R2.0.0.0 and windows NT, i get the error: 1: error -1 in extracting from zip file 2: cannot open file `bootstrap/DESCRIPTION' ALso, when i use R CMD INSTALL with the package name bootstrap.tar.gz it gives a syntax error. Can you help me with this pLease? Sincere Regards, Neba Noyan Research Assistant Center for Advanced Infrastructure and Transportation Institute Rutgers University P.S: ReLated solutions in the mailing List are below: solved : before installing using menu option, go to program files\R\rw1051\library and make a directory pubbias, then install the package from the local zip file option. 1) the zip file might have invalid internal structure. Maybe it is missing the directory `pubbias' (although the unzip code should cope with that, and that's not the error code I would expect). 2) This is not the first attempt to do something like this, and Windows had the location in use. (I have seen that one, and after a reboot it worked cleanly.) [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] edit plots from the ADE4 package
Hi again,
You could change the size font using the parameter clabel in both functions s.label
and s.class.
For the typical example in Fine, 1996 you could notice that slabel(acm$co,clabel=1)
produce different results than
slabel(acm$co,clabel=0),slabel(acm$co,clabel=0.5),slabel(acm$co,clabel=4). The size of
the point could be changed using cpoint.
Using s.class you could use col=seq(length(levels(fac))) (instead change col = rep(1,
length(levels(fac, which produces different colours according to the illustrative
factor levels.
Leonardo Trujillo
University of Southampton
Southampton Statistical Sciences Research Institute
[EMAIL PROTECTED]
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Rafael Gutierrez
Sent: 15 October 2004 14:54
To: [EMAIL PROTECTED]
Subject: [R] edit plots from the ADE4 package
Dear R-Help
i have a newbie question, how edit size fonts and colors in the plots (scatter.acm or
dudi.acm function) of multiple correspondence analysis in the ADE4 Package?
thanks?
Rafael Gutierrez
Estadístico
Unidad de Tecnología Cerro Matoso S.A.
Tel. 4-7723350 Fax. 4-7723236
*
Este correo y sus anexos pueden ser confidenciales y estar protegidos por derechos de
autor.
Estan dirigidos unica y exclusivamente para uso de el (los) destinatario(s).
Si Usted por error lo ha recibido por favor notifiquelo inmediatamente al remitente y
destruyalo de su sistema.
No debe copiar, ni imprimir, ni distribuir este correo o sus anexos, ni usarlos para
proposito alguno ni dar a conocer su contenido a persona alguna.
Las opiniones, conclusiones y otra informacion contenida en este correo no
relacionadas con el negocio oficial de Cerro Matoso, deben entenderse como personales
y de ninguna manera son avaladas por la Compañia.
*
This message and any attachments may be confidential and may...{{dropped}}
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico Hints: Balanced designs are robust to non-normality; independence (especially clustering of subjects due to systematic effects), not normality is usually the biggest real statistical problem; hypothesis tests will always reject when samples are large -- so what!; trust refers to prediction validity which has to do with study design and the validity/representativeness of the current data to future. I know that all the stats 101 tests say to test for normality, but they're full of baloney! Of course, this is free advice -- so caveat emptor! Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Building package compatible w/ R v1.9.1 and R v2.0.0?
Hi, just in the process of updating my packages for R v2.0.0 and I have not had time to followed the R v2.0.0 discussions so maybe my questions have already been answered. A concern I have is that when creating packages they should be backward compatible with R v1.9.x for a while until all users and computers has migrated to R v2.0.x. It is pretty straightforward to setup my packages so that they will *build from source* smoothly under both R v1.9.1 and R v2.0.0. But, is there an easy way to to build a binary (read Windows binary) so that it will install on R v2.0.0 as well as Rv1.9.1, or do I have to provide two seperate builds? As you already know, current status is that building under R v2.0.0 and loading in R v1.9.1 gives the error: Error in firstlib(which.lib.loc, package) : couldn't find function lazyLoad In addition: Warning message: package R.oo was built under R version 2.0.0 and building under R v1.9.1 and loading in R v2.0.0 gives the error: Error in library(R.oo) : 'R.oo' is not a valid package -- installed 2.0.0? Just to give it a try, I tried to build the package under R v2.0.0 with LazyLoad: FALSE in DESCRIPTION, but with the same error. Best wishes Henrik Bengtsson Dept. of Mathematical Statistics @ Centre for Mathematical Sciences Lund Institute of Technology/Lund University, Sweden (+2h UTC) +46 46 2229611 (off), +46 708 909208 (cell), +46 46 2224623 (fax) h b @ m a t h s . l t h . s e, http://www.maths.lth.se/~hb/ __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico 1. This was meant as a private reply so I would not roil the list. In future, when a reply takes a discussion off list, you should keep it off list, please. 2. The writings of (and personal conversations with) John Tukey and George Box are certainly primary influences, as are numerous other commentaries over the year from folks like Leo Breiman, Jerry Friedman, David Freedman, Persi Diaconis and many others. Box's original paper about robustness to non-normality was around 1952, I think, but much of what I allude to is statistical folklore, I think. Perhaps other list contributors might give you some better specific references. Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] one more Rcmdr problem
Brilliant. With R 2.0.0 patched and Rcmdr 0.9-12 all problems are gone. Thank you all, and a special thank you for the R Commander; it saves a lot of effort for teaching purposes. Best regards, Adrian -Original Message- From: Erich Neuwirth [mailto:[EMAIL PROTECTED] Sent: 15 octombrie 2004 18:27 To: Adrian Dusa Cc: [EMAIL PROTECTED] Subject: Re: [R] one more Rcmdr problem I did experience the same problem. After installing R 2.0.0 patched and downloading the source for Rcmdr_0.99-12 from John Fox's Web page http://socserv.socsci.mcmaster.ca/jfox/Misc/Rcmdr/ and recompiling the package it now works. Ctrl-C, Ctrl-X, and Ctrl-V, in the text boxes of Rcmdr now work. Adrian Dusa wrote: Hello, I'm using R 2.0.0 with the latest Rcmdr package installed from CRAN, on Windows XP Professional. When trying to copy some commands or results, either from the upper or lower text window, this causes Rcmdr to crash: R for Windows GUI front-end has encountered a problem and needs to close Did anyone have the same problem? I don't think it's my system, as it happened to reinstall my Windows just a few days ago, and the same problem occurred in the former one. Regards, Adrian Adrian Dusa Romanian Social Data Archive 1 Schitu Magureanu Bd. 050025 Bucharest sector 5 Tel. +40 21 3126618\ +40 21 3153122/ int.101 -- Erich Neuwirth, Computer Supported Didactics Working Group Visit our SunSITE at http://sunsite.univie.ac.at Phone: +43-1-4277-38624 Fax: +43-1-4277-9386 -- This message was scanned for spam and viruses by BitDefender For more information please visit http://linux.bitdefender.com/ -- This message was scanned for spam and viruses by BitDefender For more information please visit http://linux.bitdefender.com/ __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] SJava 0.65 and R 2.0.0 (slightly long)
Hi,
I had some Java code that worked with SJava 0.65 under R 1.8.*. I'm
trying to get it to work with R 2.0.0. My JVM is Sun 1.5.0 (running on
Fedora Core 2)
I downloaded and installed SJava 0.65 in my personal directory using
R CMD INSTALL SJava-0.65.tar.gz -l ~/src/Rlibrary
and before starting R I sourced RJava.bsh. However after loading the
SJava library if do javaConfig() I get
-
$classPath
[1] /omegahat/Jars/Environment.jar /..
[3] /omegahat/Jars/antlr.jar /omegahat/Jars/jas.jar
[5] /omegahat/Jars/jhall.jar
$properties
EmbeddedInR
true
InterfaceManagerClass
org/omegahat/Interfaces/NativeInterface/OmegaInterfaceManager
ForeignReferenceBaseClass
org/omegahat/R/Java/RForeignReference
java.compiler
NONE
OMEGA_HOME
OmegahatSearchPath
.,${OMEGA_HOME}/Environment/Scripts/Run,${OMEGA_HOME}/Jars/Environment.jar
java.library.path
$libraryPath
[1]
and doing .JavaInit() gives me:
JVM (nil) Env (nil)
[1] error initializing manager class Cannot find the Omegahat interface
manager class. Check you classpath!
Error in .JavaInit() : Couldn't start Java Virtual Machine: Cannot find
the Omegahat interface manager class. Check you classpath!
From the output fof javaConfig() I was'nt surprised. However I thought
that the install would set the paths to the OmegaHat interface manager.
So I set the classpath to
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/../Interfaces: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/../R/Java: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/antlr.jar: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/Environment.jar: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/jas.jar: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/jhall.jar: \
/home/rajarshi/src/Rlibrary/SJava/org/omegahat/Jars/ROmegahatExamples.jar:
and then starting up R and doing javaConfig() I get the same output as
before (except that the above paths are now in the CLASSPATH).
I also noted that the OMEGA_HOME variable is set to . This did not
seem to be right so I tried setting it to
/home/rajarshi/src/Rlibrary/SJava/org/omegahat
from my shell but on starting up R and loading SJava and running
javaConfig() I see that OMEGA_HOME is still set to
Does anybody have any pointers as to whats going on?
(I also tried SJava 0.68, but thought it installs fine and .JavaInit()
works, when I try the example on the SJava download page, my R session
crashes)
---
Rajarshi Guha [EMAIL PROTECTED] http://jijo.cjb.net
GPG Fingerprint: 0CCA 8EE2 2EEB 25E2 AB04 06F7 1BB9 E634 9B87 56EE
---
I'm related to people I don't relate to.
-Calvin
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] add-on package/windows installation probLem
You downloaded a .tar.gz file and treated it as a zip file, for some reason not in any R documentation. On Fri, 15 Oct 2004, Nebahat Noyan wrote: Hi, I have tried to download the bootstrap.tar.gz from the website below: http://cran.r-project.org/src/contrib/Old/0.50/INDEX.html I have extracted the folder to a zip file...and then use the load packages from local zip files. I use R2.0.0.0 and windows NT, i get the error: 1: error -1 in extracting from zip file 2: cannot open file `bootstrap/DESCRIPTION' ALso, when i use R CMD INSTALL with the package name bootstrap.tar.gz it gives a syntax error. Can you help me with this pLease? Perhaps that is because the file you got is for R 0.50, not R 2.0.0 (sic)? The bootstrap package has been withdrawn and does not work under R 2.0.0 AFAIK. Sincere Regards, Neba Noyan Research Assistant Center for Advanced Infrastructure and Transportation Institute Rutgers University P.S: ReLated solutions in the mailing List are below: They are *not* related. solved : before installing using menu option, go to program files\R\rw1051\library and make a directory pubbias, then install the package from the local zip file option. 1) the zip file might have invalid internal structure. Maybe it is missing the directory `pubbias' (although the unzip code should cope with that, and that's not the error code I would expect). 2) This is not the first attempt to do something like this, and Windows had the location in use. (I have seen that one, and after a reboot it worked cleanly.) [[alternative HTML version deleted]] What does the posting guide say about that? PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] add-on package/windows installation probLem
Nebahat Noyan wrote: Hi, I have tried to download the bootstrap.tar.gz from the website below: http://cran.r-project.org/src/contrib/Old/0.50/INDEX.html 0. For sure you are joking ... The package is 7.5 years old (16-Apr-1997) - too old to be installable by recent versions of R. But you might want to look into more recent packages on CRAN 1. Under Windows, you have to INSTALL via R CMD INSTALL with R = 2.0.0 (if there is no binary version available built with that version). 2. The solution in the mail given below was never the preferable one. Uwe Ligges I have extracted the folder to a zip file...and then use the load packages from local zip files. I use R2.0.0.0 and windows NT, i get the error: 1: error -1 in extracting from zip file 2: cannot open file `bootstrap/DESCRIPTION' ALso, when i use R CMD INSTALL with the package name bootstrap.tar.gz it gives a syntax error. Can you help me with this pLease? Sincere Regards, Neba Noyan Research Assistant Center for Advanced Infrastructure and Transportation Institute Rutgers University P.S: ReLated solutions in the mailing List are below: solved : before installing using menu option, go to program files\R\rw1051\library and make a directory pubbias, then install the package from the local zip file option. 1) the zip file might have invalid internal structure. Maybe it is missing the directory `pubbias' (although the unzip code should cope with that, and that's not the error code I would expect). 2) This is not the first attempt to do something like this, and Windows had the location in use. (I have seen that one, and after a reboot it worked cleanly.) [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Building package compatible w/ R v1.9.1 and R v2.0.0?
Henrik Bengtsson wrote: Hi, just in the process of updating my packages for R v2.0.0 and I have not had time to followed the R v2.0.0 discussions so maybe my questions have already been answered. A concern I have is that when creating packages they should be backward compatible with R v1.9.x for a while until all users and computers has migrated to R v2.0.x. It is pretty straightforward to setup my packages so that they will *build from source* smoothly under both R v1.9.1 and R v2.0.0. But, is there an easy way to to build a binary (read Windows binary) so that it will install on R v2.0.0 as well as Rv1.9.1, or do I have to provide two seperate builds? Yes. Same source might work, but you need separate binaries. Uwe Ligges As you already know, current status is that building under R v2.0.0 and loading in R v1.9.1 gives the error: Error in firstlib(which.lib.loc, package) : couldn't find function lazyLoad In addition: Warning message: package R.oo was built under R version 2.0.0 and building under R v1.9.1 and loading in R v2.0.0 gives the error: Error in library(R.oo) : 'R.oo' is not a valid package -- installed 2.0.0? Just to give it a try, I tried to build the package under R v2.0.0 with LazyLoad: FALSE in DESCRIPTION, but with the same error. Best wishes Henrik Bengtsson Dept. of Mathematical Statistics @ Centre for Mathematical Sciences Lund Institute of Technology/Lund University, Sweden (+2h UTC) +46 46 2229611 (off), +46 708 909208 (cell), +46 46 2224623 (fax) h b @ m a t h s . l t h . s e, http://www.maths.lth.se/~hb/ __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] one more Rcmdr problem
Dear Adrian, Since I moved the discussion of this problem to the r-devel list, I should have reported back to r-help (1) that the problem was general to tlctk (not just the Rcmdr package), and (2) that Duncan Murdoch made a change to the patched Windows version of R 2.0.0 that made the problem go away. Regards, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Adrian Dusa Sent: Friday, October 15, 2004 11:44 AM To: 'Erich Neuwirth'; 'John Fox'; 'Duncan Murdoch' Cc: [EMAIL PROTECTED] Subject: RE: [R] one more Rcmdr problem Brilliant. With R 2.0.0 patched and Rcmdr 0.9-12 all problems are gone. Thank you all, and a special thank you for the R Commander; it saves a lot of effort for teaching purposes. Best regards, Adrian -Original Message- From: Erich Neuwirth [mailto:[EMAIL PROTECTED] Sent: 15 octombrie 2004 18:27 To: Adrian Dusa Cc: [EMAIL PROTECTED] Subject: Re: [R] one more Rcmdr problem I did experience the same problem. After installing R 2.0.0 patched and downloading the source for Rcmdr_0.99-12 from John Fox's Web page http://socserv.socsci.mcmaster.ca/jfox/Misc/Rcmdr/ and recompiling the package it now works. Ctrl-C, Ctrl-X, and Ctrl-V, in the text boxes of Rcmdr now work. Adrian Dusa wrote: Hello, I'm using R 2.0.0 with the latest Rcmdr package installed from CRAN, on Windows XP Professional. When trying to copy some commands or results, either from the upper or lower text window, this causes Rcmdr to crash: R for Windows GUI front-end has encountered a problem and needs to close Did anyone have the same problem? I don't think it's my system, as it happened to reinstall my Windows just a few days ago, and the same problem occurred in the former one. Regards, Adrian Adrian Dusa Romanian Social Data Archive 1 Schitu Magureanu Bd. 050025 Bucharest sector 5 Tel. +40 21 3126618\ +40 21 3153122/ int.101 -- Erich Neuwirth, Computer Supported Didactics Working Group Visit our SunSITE at http://sunsite.univie.ac.at Phone: +43-1-4277-38624 Fax: +43-1-4277-9386 -- This message was scanned for spam and viruses by BitDefender For more information please visit http://linux.bitdefender.com/ -- This message was scanned for spam and viruses by BitDefender For more information please visit http://linux.bitdefender.com/ __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Mandrake rpm's for R 2.0.0
A binary RPM for Mandrake 10.0 official has been uploaded and will soon be available on CRAN, as well as the SRPM with which users of other flavours of Mandrake will be able to build (and hopefully contribute) their own RPM's. Please read the readme file if you have problems installing. Downloads are also possible from my homepage: http://www.webalice.it/michele.alzetta -- Michele Alzetta __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Let's see if I can get my stat 101 straight: We learned that linear regression has a set of assumptions: 1. Linearity of the relationship between X and y. 2. Independence of errors. 3. Homoscedasticity (equal error variance). 4. Normality of errors. Now, we should ask: Why are they needed? Can we get away with less? What if some of them are not met? It should be clear why we need #1. Without #2, I believe the least squares estimator is still unbias, but the usual estimate of SEs for the coefficients are wrong, so the t-tests are wrong. Without #3, the coefficients are, again, still unbiased, but not as efficient as can be. Interval estimates for the prediction will surely be wrong. Without #4, well, it depends. If the residual DF is sufficiently large, the t-tests are still valid because of CLT. You do need normality if you have small residual DF. The problem with normality tests, I believe, is that they usually have fairly low power at small sample sizes, so that doesn't quite help. There's no free lunch: A normality test with good power will usually have good power against a fairly narrow class of alternatives, and almost no power against others (directional test). How do you decide what to use? Has anyone seen a data set where the normality test on the residuals is crucial in coming up with appriate analysis? Cheers, Andy From: Federico Gherardini Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico Hints: Balanced designs are robust to non-normality; independence (especially clustering of subjects due to systematic effects), not normality is usually the biggest real statistical problem; hypothesis tests will always reject when samples are large -- so what!; trust refers to prediction validity which has to do with study design and the validity/representativeness of the current data to future. I know that all the stats 101 tests say to test for normality, but they're full of baloney! Of course, this is free advice -- so caveat emptor! Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Dear Kjetil,
I don't believe that these are BLUS residuals, but since the last n - r
effects are projections onto an orthogonal basis for the residual
subspace, they should do just fine (as long as the basis vectors have the
same length, which I think is the case, but perhaps someone can confirm).
The general idea is to transform the LS residuals into an uncorrelated,
equal-variance set.
Regards,
John
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox
-Original Message-
From: Kjetil Brinchmann Halvorsen [mailto:[EMAIL PROTECTED]
Sent: Friday, October 15, 2004 9:12 AM
To: John Fox
Cc: 'Federico Gherardini'; [EMAIL PROTECTED]
Subject: Re: [R] Testing for normality of residuals in a
regression model
John Fox wrote:
Dear Federico,
A problem with applying a standard test of normality to LS
residuals is
that the residuals are correlated and heterskedastic even if the
standard assumptions of the model hold. In a large sample, this is
unlikely to be problematic (unless there's an unusual data
configuration), but in a small sample the effect could be nontrivial.
One approach is to use BLUS residuals, which transform the
LS residuals
to a smaller set of uncorrelated, homoskedastic residuals
(assuming the
correctness of the model).
I'm not sure if this are BLUE residuals, but the following
function transform to a smaller set of independent,
homoscedastic residuals and the calls
shapiro.test:
I've proposed to make this a method for shapiro.test for lm
objects, but it is not accepted.
shapiro.test.lm
function (obj)
{
eff - effects(obj)
rank - obj$rank
df.r - obj$df.residual
if (df.r 3)
stop(To few degrees of freedom for residual for the test.)
data.name - deparse(substitute(obj))
x - eff[-(1:rank)]
res - shapiro.test(x)
res$data.name - data.name
res$method - paste(res$method, for residuals of linear model)
res
}
Kjetil
A search of R resources didn't turn up anything for BLUS, but they
shouldn't be hard to compute. This is a standard topic
covered in many
econometrics texts.
You might consider the alternative of generating a bootstrapped
confidence envelope for the QQ plot; the qq.plot() function
in the car
package will do this for a linear model.
I hope this helps,
John
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Federico
Gherardini
Sent: Friday, October 15, 2004 7:44 AM
To: [EMAIL PROTECTED]
Subject: [R] Testing for normality of residuals in a
regression model
Hi all,
Is it possible to have a test value for assessing the normality of
residuals from a linear regression model, instead of simply
relying on
qqplots?
I've tried to use fitdistr to try and fit the residuals
with a normal
distribution, but fitdsitr only returns the parameters of the
distribution and the standard errors, not the p-value. Am I missing
something?
Cheers,
Federico
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
--
Kjetil Halvorsen.
Peace is the most effective weapon of mass construction.
-- Mahdi Elmandjra
__
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Dear Federico, The problem is the same with GLS residuals -- even if the GLS transformation produces homoskedastic errors, the residuals will be correlated and heteroskedastic (with this problem tending to disappear in most instances as n grows). The central point is that residuals don't behave quite the same as errors. Regards, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Federico Gherardini Sent: Friday, October 15, 2004 11:22 AM To: [EMAIL PROTECTED] Subject: Re: [R] Testing for normality of residuals in a regression model Thank you very much for your suggestions! The residuals come from a gls model, because I had to correct for heteroscedasticity using a weighted regression... can I simply apply one of these tests (like shapiro.test) to the standardized residuals from my gls model? Cheers, Federico __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Dear Andy, At the risk of muddying the waters (and certainly without wanting to advocate the use of normality tests for residuals), I believe that your point #4 is subject to misinterpretation: That is, while it is true that t- and F-tests for regression coefficients in large sample retain their validity well when the errors are non-normal, the efficiency of the LS estimates can (depending upon the nature of the non-normality) be seriously compromised, not only absolutely but in relation to alternatives (e.g., robust regression). Regards, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Liaw, Andy Sent: Friday, October 15, 2004 11:55 AM To: 'Federico Gherardini'; Berton Gunter Cc: R-help mailing list Subject: RE: [R] Testing for normality of residuals in a regression model Let's see if I can get my stat 101 straight: We learned that linear regression has a set of assumptions: 1. Linearity of the relationship between X and y. 2. Independence of errors. 3. Homoscedasticity (equal error variance). 4. Normality of errors. Now, we should ask: Why are they needed? Can we get away with less? What if some of them are not met? It should be clear why we need #1. Without #2, I believe the least squares estimator is still unbias, but the usual estimate of SEs for the coefficients are wrong, so the t-tests are wrong. Without #3, the coefficients are, again, still unbiased, but not as efficient as can be. Interval estimates for the prediction will surely be wrong. Without #4, well, it depends. If the residual DF is sufficiently large, the t-tests are still valid because of CLT. You do need normality if you have small residual DF. The problem with normality tests, I believe, is that they usually have fairly low power at small sample sizes, so that doesn't quite help. There's no free lunch: A normality test with good power will usually have good power against a fairly narrow class of alternatives, and almost no power against others (directional test). How do you decide what to use? Has anyone seen a data set where the normality test on the residuals is crucial in coming up with appriate analysis? Cheers, Andy From: Federico Gherardini Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico Hints: Balanced designs are robust to non-normality; independence (especially clustering of subjects due to systematic effects), not normality is usually the biggest real statistical problem; hypothesis tests will always reject when samples are large -- so what!; trust refers to prediction validity which has to do with study design and the validity/representativeness of the current data to future. I know that all the stats 101 tests say to test for normality, but they're full of baloney! Of course, this is free advice -- so caveat emptor! Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] fidelity of generated raster images (R and perl)
Wrap-up: Different solutions were proposed for turning a matrix into a bitmap. Henrik Bengtsson [EMAIL PROTECTED] R.classes bundle from (http://www.maths.lth.se/help/R/R.classes/) MonochromeImage R.classes is still not compatible with 2.0.0 (minor modification to DESCRIPTION etc are needed) so you have use R v1.9.1 to try the below. img - MonochromeImage(1-bits) # 1=white, 0=black image(img) plot(img) persp(img, phi=60, theta=150) write(img, foo.pbm) # Write directly Martin Maechler [EMAIL PROTECTED] I think the most typical alternative is to use image() [EMAIL PROTECTED] Maybe use the write.pnm() function in the pixmap package Another package you could explore is rimage, which looks relevant. For the sake of being conventional and minimizing dependencies, I had good success with the image() function. It is very nice to know there are other approaches if I need to optimize/customize what I am doing more for this. (As usual) learned a few more things about 'R' from the examples provided by others. Most grateful for your help. Regards, Scott __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] Testing for normality of residuals in a regression model
Hi John, Your point is well taken. I was only thinking about the shape of the distribution, and neglected the cases of, say, symmetric long tailed distributions. However, I think I'd still argue that other tools are probably more useful than normality tests (e.g., robust methods, as you mentioned). To take the point a bit further, let's say we test for normality and it's rejected. What do we do then? Well, if the non-normality is caused by outliers, we can try robust methods. If not, what do we do? We can try to see if some sort of transformation would bring the residuals closer to normally distributed, but if the interest is in inference on the coefficients, those inferences on the `final' model are potentially invalid. What's one to do then? Also, I was told by someone very smart that fitting OLS to data with heteroscedastic errors can make the residuals look `more normal' than they really are... Don't know how true that is, though. Best, Andy From: John Fox Dear Andy, At the risk of muddying the waters (and certainly without wanting to advocate the use of normality tests for residuals), I believe that your point #4 is subject to misinterpretation: That is, while it is true that t- and F-tests for regression coefficients in large sample retain their validity well when the errors are non-normal, the efficiency of the LS estimates can (depending upon the nature of the non-normality) be seriously compromised, not only absolutely but in relation to alternatives (e.g., robust regression). Regards, John John Fox Department of Sociology McMaster University Hamilton, Ontario Canada L8S 4M4 905-525-9140x23604 http://socserv.mcmaster.ca/jfox -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Liaw, Andy Sent: Friday, October 15, 2004 11:55 AM To: 'Federico Gherardini'; Berton Gunter Cc: R-help mailing list Subject: RE: [R] Testing for normality of residuals in a regression model Let's see if I can get my stat 101 straight: We learned that linear regression has a set of assumptions: 1. Linearity of the relationship between X and y. 2. Independence of errors. 3. Homoscedasticity (equal error variance). 4. Normality of errors. Now, we should ask: Why are they needed? Can we get away with less? What if some of them are not met? It should be clear why we need #1. Without #2, I believe the least squares estimator is still unbias, but the usual estimate of SEs for the coefficients are wrong, so the t-tests are wrong. Without #3, the coefficients are, again, still unbiased, but not as efficient as can be. Interval estimates for the prediction will surely be wrong. Without #4, well, it depends. If the residual DF is sufficiently large, the t-tests are still valid because of CLT. You do need normality if you have small residual DF. The problem with normality tests, I believe, is that they usually have fairly low power at small sample sizes, so that doesn't quite help. There's no free lunch: A normality test with good power will usually have good power against a fairly narrow class of alternatives, and almost no power against others (directional test). How do you decide what to use? Has anyone seen a data set where the normality test on the residuals is crucial in coming up with appriate analysis? Cheers, Andy From: Federico Gherardini Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico Hints: Balanced designs are robust to non-normality; independence (especially clustering of subjects due to systematic effects), not normality is usually the biggest real statistical problem; hypothesis tests will always reject when samples are large -- so what!; trust refers to prediction validity which has to do with study design and the validity/representativeness of the current data to future. I know that all the stats 101 tests say to test for normality, but they're full of baloney! Of course, this is free advice -- so caveat emptor! Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide!
Re: [R] Testing for normality of residuals in a regression model
OK, I'll expose myself: I tend to do normal probability plots of residuals (usely deletion / studentized residuals as described by Venables and Ripley in Modern Applied Statistics with S, 4th ed, MASS4). If the plots look strange, I do something. I'll check apparent outliers for coding and data entry errors, and I often delete those points from the analysis even if I can't find a reason why. Robust regression will usually handle this type of problem, and I am gradually migrating to increasing use of robust regression, especially the procedures recommended by MASS4. . However, I recently encountered a situation that would be masked by standard use of robust regression without examining residual plots: A normal probability plot looked like three parallel straight lines with gaps, suggesting a mixture of 3 normal distributions with different means and a common standard deviation. Further investigation revealed that an important 3-level explanatory variable that had been miscoded. When this was corrected, that variable entered the model and the gaps in the normal plot disappeared. I tend NOT to use tests of normality for the reasons Andy mentioned. Instead, I do various kinds of diagnostic plots and modify my model or investigate the data in response to what I see. Comments? hope this helps. spencer graves Liaw, Andy wrote: Let's see if I can get my stat 101 straight: We learned that linear regression has a set of assumptions: 1. Linearity of the relationship between X and y. 2. Independence of errors. 3. Homoscedasticity (equal error variance). 4. Normality of errors. Now, we should ask: Why are they needed? Can we get away with less? What if some of them are not met? It should be clear why we need #1. Without #2, I believe the least squares estimator is still unbias, but the usual estimate of SEs for the coefficients are wrong, so the t-tests are wrong. Without #3, the coefficients are, again, still unbiased, but not as efficient as can be. Interval estimates for the prediction will surely be wrong. Without #4, well, it depends. If the residual DF is sufficiently large, the t-tests are still valid because of CLT. You do need normality if you have small residual DF. The problem with normality tests, I believe, is that they usually have fairly low power at small sample sizes, so that doesn't quite help. There's no free lunch: A normality test with good power will usually have good power against a fairly narrow class of alternatives, and almost no power against others (directional test). How do you decide what to use? Has anyone seen a data set where the normality test on the residuals is crucial in coming up with appriate analysis? Cheers, Andy From: Federico Gherardini Berton Gunter wrote: Exactly! My point is that normality tests are useless for this purpose for reasons that are beyond what I can take up here. Thanks for your suggestions, I undesrtand that! Could you possibly give me some (not too complicated!) links so that I can investigate this matter further? Cheers, Federico Hints: Balanced designs are robust to non-normality; independence (especially clustering of subjects due to systematic effects), not normality is usually the biggest real statistical problem; hypothesis tests will always reject when samples are large -- so what!; trust refers to prediction validity which has to do with study design and the validity/representativeness of the current data to future. I know that all the stats 101 tests say to test for normality, but they're full of baloney! Of course, this is free advice -- so caveat emptor! Cheers, Bert __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Spencer Graves, PhD, Senior Development Engineer O: (408)938-4420; mobile: (408)655-4567 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] categorical varibles in coxph
Hello, I wonder when I do coxph in R: coxph( Surv(start, stop, event) ~ x, data=test) If x is a categorical varible (1,2,3,4,5), should I creat four dummy varibles for it? if yes, how can I get the overall p value on x other than for each dummy variable? Thanks Lisa Wang Princess Margaret Hospital Phone 416 946 4501 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Using accrualReport.s in Hmisc/Design
Hello, I apologize for the post if this is not the appropriate place. I have searched the R List archives and the Vanderbilt.edu website, but am not able to understand how to use the accrualReport.s and associated scripts available at the Vanderbilt site. It must be very simple, but I am at a loss for ideas. Any suggestions or pointers in the direction of examples? thanks very much. aric __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Using accrualReport.s in Hmisc/Design
Aric Gregson wrote: Hello, I apologize for the post if this is not the appropriate place. I have searched the R List archives and the Vanderbilt.edu website, but am not able to understand how to use the accrualReport.s and associated scripts available at the Vanderbilt site. It must be very simple, but I am at a loss for ideas. Any suggestions or pointers in the direction of examples? thanks very much. aric __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html That is part of a new package for clinical trials reporting that is not packaged for use by 'outsiders' yet. The R package name will be rreport. -- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] categorical varibles in coxph
On Fri, 15 Oct 2004, Lisa Wang wrote: Hello, I wonder when I do coxph in R: coxph( Surv(start, stop, event) ~ x, data=test) If x is a categorical varible (1,2,3,4,5), should I creat four dummy varibles for it? if yes, how can I get the overall p value on x other than for each dummy variable? No. Use coxph( Surv(start, stop, event) ~ factor(x), data=test) or define x as a factor. For an overal test use anova(): eg data(pbc) model-coxph(Surv(time,status)~factor(edtrt)+bili, data=pbc) model Call: coxph(formula = Surv(time, status) ~ factor(edtrt) + bili, data = pbc) coef exp(coef) se(coef)z p factor(edtrt)0.5 0.629 1.88 0.2297 2.74 6.2e-03 factor(edtrt)1 1.664 5.28 0.2762 6.02 1.7e-09 bili 0.119 1.13 0.0129 9.29 0.0e+00 Likelihood ratio test=127 on 3 df, p=0 n= 418 anova(model) Analysis of Deviance Table Cox model: response is Surv(time, status) Terms added sequentially (first to last) Df Deviance Resid. Df Resid. Dev NULL 4181746.94 factor(edtrt) 262.13 4161684.82 bili165.11 4151619.71 summary(model) Call: coxph(formula = Surv(time, status) ~ factor(edtrt) + bili, data = pbc) n= 418 coef exp(coef) se(coef)z p factor(edtrt)0.5 0.629 1.88 0.2297 2.74 6.2e-03 factor(edtrt)1 1.664 5.28 0.2762 6.02 1.7e-09 bili 0.119 1.13 0.0129 9.29 0.0e+00 exp(coef) exp(-coef) lower .95 upper .95 factor(edtrt)0.5 1.88 0.533 1.20 2.94 factor(edtrt)15.28 0.189 3.07 9.07 bili 1.13 0.887 1.10 1.16 Rsquare= 0.262 (max possible= 0.985 ) Likelihood ratio test= 127 on 3 df, p=0 Wald test= 193 on 3 df, p=0 Score (logrank) test = 281 on 3 df, p=0 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] tree version 1.0-16
Kalaylioglu, Zeynep (IMS) wrote: Hi There is something weird going on with the tree package version 1.0-16. So I want to download an updated tree package. I found that R website has version 1.0-18. I want to try this one and see how the tree algorithm is performing. What is the best way to download tree package version 1.0-18 which can be found in index of src/contrib at www.r-project.org? So far I tried two different ways none of which has worked for me: 1) I tried to download it by running the command install.packages(tree,C:\Program Files\R\rw1081\library,http://cran.r-project.org/src/contrib,destdir= C:\Program http://cran.r-project.org/src/contrib,destdir=C:/Program Files\R\rw1081\library,installWithVers=TRUE) But it tries to connect to http://cran.r-project.org/src/contrib/bin/windows/contrib/1.8/PACKAGES http://cran.r-project.org/src/contrib/bin/windows/contrib/1.8/PACKAGES instead of the URL I enter above. How about Packages -- Install package(s) from CRAN...? You are making things unnecessarily difficult. 2) Also I tried to download it by copying the zip file first tree_1.0-18.tar.gz Zip files don't end with .tar.gz! These are tarred and gzipped source packages that you are almost certainly not set up to install. Your best bet is to try the simplest things first, which on Windows is to use the Packages menu. Best, Jim http://cran.r-project.org/src/contrib/tree_1.0-18.tar.gz in my directory and then using the Packages-Install package from local -zip files in the R menu. How can I download this version of tree? Thanks. Zeynep [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- James W. MacDonald Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
RE: [R] combine many .csv files into a single file/data frame
On 15-Oct-04 bogdan romocea wrote: Dear R users, I have a few hundred .csv files which I need to put together (full outer joins on a common variable) to do a factor analysis. Each file may contain anywhere from a few hundred to a few thousand rows. What would be the most efficient way to do this in R? Please include some sample code if applicable. If you're using Linux/Unix, you should consider using the 'join' command. Simple example (jpoining files j1 and j2): $ cat j1 1 a 1 b 2 c 2 d 2 e 3 f 3 g 4 h $ cat j2 1 A 1 B 2 C 3 D 3 E 3 F $ join j1 j2 1 a A 1 a B 1 b A 1 b B 2 c C 2 d C 2 e C 3 f D 3 f E 3 f F 3 g D 3 g E 3 g F See 'man join' for details of options which you can use to adapt the command to your needs. Hoping this helps, Ted. E-Mail: (Ted Harding) [EMAIL PROTECTED] Fax-to-email: +44 (0)870 094 0861 [NB: New number!] Date: 15-Oct-04 Time: 21:41:44 -- XFMail -- __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] power in a specific frequency band
Dear R users I have a really simple question (hoping for a really simple answer :-): Having estimated the spectral density of a time series x (heart rate data) with: x.pgram - spectrum(x,method=pgram) I would like to compute the power in a specific energy band. Assuming that frequency(x)=4 (Hz), and that I am interested in the band between f1 and f2, is the power in the band simply the following? sum(x.pgram$spec[(x.pgram$freq f1/frequency(x)) (x.pgram$freq = f2/frequency(x))]) If it is so, are the returned units the same units as the original time series, but squared (if x is bpm, then the power is in bpm^2)? I own a copy of Venables and Ripley (MASS 2003), but I was not able to extract this information from the time series chapter thanks for any help (please cc to my e-mail, if possible) giuseppe Giuseppe Pagnoni Dept. Psychiatry and Behavioral Sciences Emory University 1639 Pierce Drive, Suite 4000 Atlanta, GA, 30322 tel: 404.712.8431 fax: 404.727.3233 __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Testing for normality of residuals in a regression model
Liaw, Andy wrote: . . . . Also, I was told by someone very smart that fitting OLS to data with heteroscedastic errors can make the residuals look `more normal' than they really are... Don't know how true that is, though. Certainly true, since the residuals will be a kind of average, so the CLT works. (Think that is in Seber, Linear Regression Analysis, 1977) Kjetil Best, Andy -- Kjetil Halvorsen. Peace is the most effective weapon of mass construction. -- Mahdi Elmandjra __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Problem Compiling R-2.0.0 on Linux Alpha
Hi folks: I'm trying to compile R-2.0.0 on my Alpha (164SX; alphapca56-unknown-linux-gnu) running RedHat Linux 7.2 with gcc3.4.1. Here is where the compile terminates: building package 'tcltk' mkdir -p -- ../../../library/tcltk/R mkdir -p -- ../../../library/tcltk/demo mkdir -p -- ../../../library/tcltk/exec mkdir -p -- ../../../library/tcltk/man make[4]: Entering directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/t cltk/src' making init.d from init.c making tcltk.d from tcltk.c make[5]: Entering directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/t cltk/src' make[5]: `Makedeps' is up to date. make[5]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/tc ltk/src' make[5]: Entering directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/t cltk/src' gcc -I../../../../include -I/usr/local/tcl8.4.7/include -I/usr/local/tk8.4.7/inc lude -I/usr/X11R6/include -I/usr/local/include -mieee-with-inexact -fPIC -g -O 2 -c init.c -o init.o gcc -I../../../../include -I/usr/local/tcl8.4.7/include -I/usr/local/tk8.4.7/inc lude -I/usr/X11R6/include -I/usr/local/include -mieee-with-inexact -fPIC -g -O 2 -c tcltk.c -o tcltk.o gcc -shared -L/usr/local/lib -o tcltk.so init.o tcltk.o -L/usr/local/tcl8.4.7/l ib -ltcl8.4 -L/usr/local/tk8.4.7/lib -ltk8.4 -L/usr/X11R6/lib -lX11 -ldl -lieee -lm -lm -L../../../../lib -lR mkdir -p -- ../../../../library/tcltk/libs make[5]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/tc ltk/src' make[4]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/tc ltk/src' make[3]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library/tc ltk' Error in names-.default(`*tmp*`, value = c(R, Platform, Date, : names attribute [4] must be the same length as the vector [3] Execution halted make[2]: *** [R] Error 1 make[2]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src/library' make[1]: *** [R] Error 1 make[1]: Leaving directory `/home1/rajiv/software/src-7.2/R-2.0.0/src' make: *** [R] Error 1 Does this indicate a problem with R, or something in my system installation? A search on R help list archives did not turn up this particular error. Any help is much appreciate. Rajiv Rajiv Prasad Scientist, Hydrology Group Pacific Northwest National Laboratory, P.O. Box 999, MSIN K9-33 Richland, WA 99352 Voice: (509) 375-2096 Fax: (509) 372-6089 Email: [EMAIL PROTECTED] [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] 20,000 * 6 data values
Hello, Rusers: What is the maximum number of data R can handle? Or I have to use SAS? I am trying to do some microarray data analysis. But I am totally new. Did anyone use R to do microarray analysis? Many thanks, Sun __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] sapply and loop
Dear all, I am doing 200 times simulation. For each time, I generate a matrix and define some function on this matrix to get a 6 dimension vector as my results. As the loop should be slow, I generate 200 matrice first, and save them into a list named ma, then I define zz-sapply(ma, myfunction) To my surprise, It almost costs me the same time to get my results if I directly use a loop from 1 to 200. Is it common? Can I improve any further? Ps, how to count the exact time to finish my code? Thanks. Zhen __ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
