Re: [R] Bug in formals-
Frank E Harrell Jr [EMAIL PROTECTED] writes: Deepayan Sarkar wrote: On 9/26/06, Frank E Harrell Jr [EMAIL PROTECTED] wrote: I think this is new since a previous version of R: h - function(x, trantab) trantab[x] w - 6:4 names(w) - c('cat','dog','giraffe') w cat dog giraffe 6 5 4 formals(h) - list(x=numeric(0), trantab=w) h function (x = numeric(0), trantab = c(6, 5, 4)) trantab[x] You can see that the names have been dropped from trantab's default values. Are you sure? I get formals(h) $x numeric(0) $trantab cat dog giraffe 6 5 4 h(1) cat 6 R version 2.4.0 beta (2006-09-21 r39463) x86_64-unknown-linux-gnu -Deepayan Deepayan - You are correct. h('cat') is 6 as intended. I just looked at the function definition - the names attribute doesn't show for some reason. I was expecting function(..., trantab=c(cat=6, ..). It's a generic deparsing issue. Also: substitute(f(w),list(w=w)) f(c(6, 5, 4)) We should probably fix it at some point. Notice though that you can't parse your way into such situations, only get there by computing on the language. -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
[R] multiple imputation
Hi, is it correct that multiple-Imputation like mice http://www.imputation.com can't understand as a standard data-mining task, beacuse i haven't a generalization mechanism perform the model on complete new and bigger dataset with a predict method!? many thanks regards, christian __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] New project: littler for GNU R
Jeffrey Horner [EMAIL PROTECTED] writes: Seth Falcon wrote: Wow, looks neat. OS X users will be unhappy with your naming choice as the default filesystem there is not case-sensitive :-( IOW, r and R do the same thing. I would expect it to otherwise work on OS X so a change of some sort might be worthwhile. (I'm always amazed at how I can miss the simplest details. I probably knew at some point that OS X shipped with a case-sensitive file system, which you can turn off somehow, but forgot. Thank goodness for peer review.) littler will install into /usr/local/bin by default, so I don't think there's a clash with the Mac binary provided by CRAN, right? It depends what you mean by clash :-) If both are on the PATH, then you get the first one, I suspect, when running either 'R' or 'r'. I haven't tested this bit yet, but on my OS X laptop I can invoke a new R session using either 'R' or 'r' (using an R built from source, not the R GUI app thingie). So IMO, a different name or an integration into the R script in some way would be a big improvement. 'r' is cute, but going down the road of tools with the same name except for caps leads to confusion (for me). For example, R CMD build/INSTALL still catches me up after a number of years. + seth __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] exact 95% confidence intervals
Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Thanks a lot! My best! -- *** Xin Meng Capitalbio Corporation National Engineering Research Center for Beijing Biochip Technology BioPharma-informatics Software Dept. Research Engineer Tel: +86-10-80715888/80726868-6438 Fax: +86-10-80726790 [EMAIL PROTECTED] Address:18 Life Science Parkway, Changping District, Beijing 102206, China __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] How to pass expression as an argument
Hi, I am writing a function and need to pass a function expression as an argument, for instance, myfun - function( express) { x- c(1,2,3) y-express } if I call the above function by myfun( x*2 ), I get 2 as the result, instead of 2,4,6 , could someone help me to fix this problem ? Furthermore, is that possible to operate this expression on different variables? for example, in myfun(), I might want to get u*2, z*2, etc. without having to say x-u, x-z to match the expression. Thanks a lot for your help. best __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] How to pass expression as an argument
Tong Wang [EMAIL PROTECTED] writes: Hi, I am writing a function and need to pass a function expression as an argument, for instance, myfun - function( express) { x- c(1,2,3) y-express } if I call the above function by myfun( x*2 ), I get 2 as the result, instead of 2,4,6 , could someone help me to fix this problem ? Furthermore, is that possible to operate this expression on different variables? for example, in myfun(), I might want to get u*2, z*2, etc. without having to say x-u, x-z to match the expression. Thanks a lot for your help. Either pass the expression explicitly using myfun(quote(x*2)) or myfun(expression(x*2)) or use substitute(express) inside the function. -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
Re: [R] Accessing C- source code of R
That's exactly what I was looking for... Thanks! Gunther -Ursprüngliche Nachricht- Von: Uwe Ligges [mailto:[EMAIL PROTECTED] Gesendet: Dienstag, 26. September 2006 10:13 An: Gunther Höning Cc: r-help@stat.math.ethz.ch Betreff: Re: [R] Accessing C- source code of R Gunther Höning wrote: Dear list, I'm looking for the sources code of parts of R, (e.g. spline). Does anyone know where I can access it ? I plan to write a corresponding R Help Desk article on Accessing the source. A draft is available from: http://www.statistik.uni-dortmund.de/~ligges/R_Help_Desk_preview.pdf Can you please tell me if this description is sufficient? Thanks, Uwe Ligges Gunther __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exact 95% confidence intervals
XinMeng [EMAIL PROTECTED] writes: Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Just run binom.test on the data from each column (or row, depending on which way you turn the tables). -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
[R] Impossible to merge with a zero rows data frame?
I'm trying to merge two data frames. One of them is a zero rows data frame. I'm using the merge parameter 'all.x = TRUE' so I'd expect to obtain all the rows of x. In fact the merge help says: all.x: logical; if 'TRUE', then extra rows will be added to the output, one for each row in 'x' that has no matching row in 'y'. These rows will have 'NA's in those columns that are usually filled with values from 'y'. To let you test the problem here is some code: L3 - LETTERS[1:3] (d - data.frame(cbind(x=1, y=1), fac=sample(L3, 1, repl=TRUE))) e - d[-1,] so now we have that: e [1] x y fac 0 rows (or 0-length row.names) here is the merge: merge(d, e, by.x = c(x), by.y = c(x), all.x = TRUE) I'd expect something like: x y.x fac.x y.y fac.y 1 1 1 B NANA instead of: Error in merge.data.frame(d, e, by.x = c(x), by.y = c(x), all.x = TRUE) : no row to correspond (I'm traslating the error message, so it could be a little different) My questions are: Is there a way to obtain my desired result? In the context in which I'd like to use the code the second data frame is the result of a query and a lot of overwork would be added if I have to check if the dataframe is a zero rows one BEFORE the merge (in fact I do a sequence of several merge) Is this behaviour of the merge command correct? Why did the developers choose it? P.S.: I tested this code both in R 2.3.0 and in 2.3.1 Thank you in advance. Sandro Bonfigli __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Any hot-deck imputation packages?
Hi I found on google that there is an implementation of hot-deck imputation in SAS: http://ideas.repec.org/c/boc/bocode/s366901.html Is there anything similar in R? Many Thanks Eleni Rapsomaniki __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Any hot-deck imputation packages?
I'm not sure if this is exactly what you're after, but it might be worth looking at aregImpute in the Hmisc package. On 27/09/06, Eleni Rapsomaniki [EMAIL PROTECTED] wrote: Hi I found on google that there is an implementation of hot-deck imputation in SAS: http://ideas.repec.org/c/boc/bocode/s366901.html Is there anything similar in R? Many Thanks Eleni Rapsomaniki __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- = David Barron Said Business School University of Oxford Park End Street Oxford OX1 1HP [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Constrained OLS regression
Hello R helpers, I am trying to do a linear OLS regression of y on two variables x1 and x2. I want to constrain the coefficients of x1 and x2 to sum up to 1. and therefore run a constrained OLS. Can anybody help with this? (I have seen some answers to similar questions but it was not clear to me what I need to do) - I have tried the lm function with offset but I must not have used it properly. Thanks, Spyros __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Constrained OLS regression
Have a look at the linear.hypothesis function in the car package. For example: mod.duncan - lm(prestige ~ income + education, data=Duncan) linear.hypothesis(mod.duncan, income + education = 1) Linear hypothesis test Hypothesis: income + education = 1 Model 1: prestige ~ income + education Model 2: restricted model Res.DfRSS Df Sum of Sq F Pr(F) 1 42 7506.7 2 43 8045.2 -1-538.5 3.0129 0.08994 . --- Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 On 27/09/06, Mesomeris, Spyros [CIR] [EMAIL PROTECTED] wrote: Hello R helpers, I am trying to do a linear OLS regression of y on two variables x1 and x2. I want to constrain the coefficients of x1 and x2 to sum up to 1. and therefore run a constrained OLS. Can anybody help with this? (I have seen some answers to similar questions but it was not clear to me what I need to do) - I have tried the lm function with offset but I must not have used it properly. Thanks, Spyros __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- = David Barron Said Business School University of Oxford Park End Street Oxford OX1 1HP __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] New project: littler for GNU R
Gabor Grothendieck wrote: I think this is quoted out of context. I was referring to Duncan's post which shows an example of piping R code. No, that was Jeffrey's post that showed the pipe. All that was in my post was a question about why this isn't a patch to the standard R script. I don't think anyone has answered that, so I assume the reason is that it just seemed easier to write a separate script than to change the standard one. If someone else designs, documents and commits the necessary changes to the standard script, I'd be happy to follow suit with the Windows R.exe. Duncan Murdoch On 9/26/06, Dirk Eddelbuettel [EMAIL PROTECTED] wrote: On 26 September 2006 at 22:17, Gabor Grothendieck wrote: | The real problem is that one wants to pipe the data in, not the | R source. The idea is that one successively transforms the | data in successive elements of the pipeline. But that is what our filesize example does:: | On 9/26/06, Duncan Murdoch [EMAIL PROTECTED] wrote: | On 9/26/2006 1:04 PM, Jeffrey Horner wrote: [...] | But unlike bc(1), GNU R has a vast number of statistical | functions. For example, we can quickly compute a summary() and show | a stem-and-leaf plot for file sizes in a given directory via | |$ ls -l /boot | awk '!/^total/ {print $5}' | \ | r -e 'fsizes - as.integer(readLines()); |print(summary(fsizes)); stem(fsizes)' | Min. 1st Qu. MedianMean 3rd Qu.Max. | 13 512 110100 486900 768400 4735000 |Loading required package: grDevices | | The decimal point is 6 digit(s) to the right of the | | | 0 | 002223 | 0 | 5557778899 | 1 | 112233 | 1 | 5 | 2 | | 2 | | 3 | | 3 | | 4 | | 4 | 7 Data to be processed on stdin, command via -e 'some long expression'. To make it simpler, here is a somewhat useless example of r piping into r (which I've indented for readability): $ r -e 'set.seed(42); sapply(rnorm(5),function(x) cat(x,\n))' | \ r -e 'cat(sum(abs(as.numeric(readLines(, \n)' 3.335916 Isn't that something where, to quote you, one wants to pipe the data in, not the R source ? Dirk -- Hell, there are no rules here - we're trying to accomplish something. -- Thomas A. Edison __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] New project: littler for GNU R
Jeffrey Horner wrote: Seth Falcon wrote: Jeffrey Horner [EMAIL PROTECTED] writes: [...] littler will install into /usr/local/bin by default, so I don't think there's a clash with the Mac binary provided by CRAN, right? It depends what you mean by clash :-) If both are on the PATH, then you get the first one, I suspect, when running either 'R' or 'r'. I haven't tested this bit yet, but on my OS X laptop I can invoke a new R session using either 'R' or 'r' (using an R built from source, not the R GUI app thingie). Good point, but the executable path can be named absolutely in hash-bang scripts. Relative paths work as well with the use of '/usr/bin/env program' as is described in the littler announcement, but then you don't get to pass arguments to 'program', just to the hash-bang script. So IMO, a different name or an integration into the R script in some way would be a big improvement. But I'd like to know why there's an R script in the first place. Why not just an executable as on windows? There need to be (at least) two separate executables: on Windows they are R.exe which invokes Rterm.exe. The reason for this is that R CMD foo invokes the script foo, it doesn't run Rterm at all. On Unix-alikes the first is a script rather than an exe. The reason the first is a script on Unix is that whoever wrote it preferred the sh script language to C for handling the things it needs to do, but that's not generally available in Windows, so we were stuck with C. Duncan Murdoch 'r' is cute, but going down the road of tools with the same name except for caps leads to confusion (for me). For example, R CMD build/INSTALL still catches me up after a number of years. That's a different problem than case-sensitivity. The word 'build' must have had a different semantic than INSTALL, and I'm not sure why one was all caps and the other isn't. Jeff __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Constrained OLS regression
you could reparameterize, e.g., x1 - runif(100, -4, 4) x2 - runif(100, -4, 4) X - cbind(1, x1 , x2) y - rnorm(100, as.vector(X %*% c(5, -3, 4)), 2) ## fn - function(betas){ betas - c(betas, 1 - betas[2]) crossprod(y - X %*% betas)[1, ] } opt - optim(c(5, -3), fn, method = BFGS) c(opt$par, 1 - opt$par[2]) 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/(0)16/336899 Fax: +32/(0)16/337015 Web: http://med.kuleuven.be/biostat/ http://www.student.kuleuven.be/~m0390867/dimitris.htm - Original Message - From: Mesomeris, Spyros [CIR] [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Sent: Wednesday, September 27, 2006 12:51 PM Subject: [R] Constrained OLS regression Hello R helpers, I am trying to do a linear OLS regression of y on two variables x1 and x2. I want to constrain the coefficients of x1 and x2 to sum up to 1. and therefore run a constrained OLS. Can anybody help with this? (I have seen some answers to similar questions but it was not clear to me what I need to do) - I have tried the lm function with offset but I must not have used it properly. Thanks, Spyros __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] How to pass expression as an argument
On 27 Sep 2006 10:57:15 +0200, Peter Dalgaard [EMAIL PROTECTED] wrote: Tong Wang [EMAIL PROTECTED] writes: Hi, I am writing a function and need to pass a function expression as an argument, for instance, myfun - function( express) { x- c(1,2,3) y-express } if I call the above function by myfun( x*2 ), I get 2 as the result, instead of 2,4,6 , could someone help me to fix this problem ? Furthermore, is that possible to operate this expression on different variables? for example, in myfun(), I might want to get u*2, z*2, etc. without having to say x-u, x-z to match the expression. Thanks a lot for your help. Either pass the expression explicitly using myfun(quote(x*2)) or myfun(expression(x*2)) or use substitute(express) inside the function. Regarding the second option see the source of curve by entering: curve __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] New project: littler for GNU R
On 9/27/2006 7:06 AM, Duncan Murdoch wrote: Gabor Grothendieck wrote: I think this is quoted out of context. I was referring to Duncan's post which shows an example of piping R code. No, that was Jeffrey's post that showed the pipe. All that was in my post was a question about why this isn't a patch to the standard R script. I don't think anyone has answered that, so I assume the reason is that it just seemed easier to write a separate script than to change the standard one. I've been reminded offline that this was discussed a few months ago by R Core, and some tricky questions were raised. Nobody had the time to address them before 2.4.0, but I think the hope is that we'll have time after it is released to nail down the decisions on this so that something makes it into 2.5.0. Expect some more comments on the tricky questions, but not for a couple of weeks: 2.4.0 is due out in 6 days, and that's the priority right now. Duncan Murdoch If someone else designs, documents and commits the necessary changes to the standard script, I'd be happy to follow suit with the Windows R.exe. Duncan Murdoch On 9/26/06, Dirk Eddelbuettel [EMAIL PROTECTED] wrote: On 26 September 2006 at 22:17, Gabor Grothendieck wrote: | The real problem is that one wants to pipe the data in, not the | R source. The idea is that one successively transforms the | data in successive elements of the pipeline. But that is what our filesize example does:: | On 9/26/06, Duncan Murdoch [EMAIL PROTECTED] wrote: | On 9/26/2006 1:04 PM, Jeffrey Horner wrote: [...] | But unlike bc(1), GNU R has a vast number of statistical | functions. For example, we can quickly compute a summary() and show | a stem-and-leaf plot for file sizes in a given directory via | |$ ls -l /boot | awk '!/^total/ {print $5}' | \ | r -e 'fsizes - as.integer(readLines()); |print(summary(fsizes)); stem(fsizes)' | Min. 1st Qu. MedianMean 3rd Qu.Max. | 13 512 110100 486900 768400 4735000 |Loading required package: grDevices | | The decimal point is 6 digit(s) to the right of the | | | 0 | 002223 | 0 | 5557778899 | 1 | 112233 | 1 | 5 | 2 | | 2 | | 3 | | 3 | | 4 | | 4 | 7 Data to be processed on stdin, command via -e 'some long expression'. To make it simpler, here is a somewhat useless example of r piping into r (which I've indented for readability): $ r -e 'set.seed(42); sapply(rnorm(5),function(x) cat(x,\n))' | \ r -e 'cat(sum(abs(as.numeric(readLines(, \n)' 3.335916 Isn't that something where, to quote you, one wants to pipe the data in, not the R source ? Dirk -- Hell, there are no rules here - we're trying to accomplish something. -- Thomas A. Edison __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multiple imputation
[EMAIL PROTECTED] wrote: Hi, is it correct that multiple-Imputation like mice http://www.imputation.com can't understand as a standard data-mining task, beacuse i haven't a generalization mechanism perform the model on complete new and bigger dataset with a predict method!? many thanks regards, christian This is something we need. I have not written a predict method for aregImpute in the Hmisc package yet (and soon a completely re-written version of aregImpute will be posted) but the framework in aregImpute may allow such a method to be written. Volunteers welcome. Frank -- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exact 95% confidence intervals
Peter Dalgaard wrote: XinMeng [EMAIL PROTECTED] writes: Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Just run binom.test on the data from each column (or row, depending on which way you turn the tables). But beware of exact binomial intervals - they are often not very accurate. Wilson and other intervals are generally better. For example see the binconf function in the Hmisc package. -- Frank E Harrell Jr Professor and Chair School of Medicine Department of Biostatistics Vanderbilt University __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] equivalent of model.tables for an lm.object?
Dear all, I run a linear model with three significant explanatory variabels x1: a factor with 4 levels x2 and x3: factors with two levels each x4: continuous model - lm(y ~ x1 + x2 * x3 + x4) The data is not perfectly balanced between the different factor-combinations and I use treatment contrasts. With an aov.object, I assume I could have used model.tables(aov.object, type = means, se = TRUE), to get the means and se for all factor combinations. In an lm.object like mine, I calculate the means 'manually' from the Estimates (for sure it could be done with a script, but fair enough). For the standard error of the means, I started out using formulas of a variance of a sum of two variables, but I messed things up with the interaction. Is there a way to calculate the standard error of the means from Estimates and Std.Error (or other information) from the lm.object? Thanks in advance for any advice! Best regards, Henrik -- Henrik Pärn Department of Biology NTNU 7491 Trondheim Norway +47 735 96282 (office) +47 909 89 255 (mobile) +47 735 96100 (fax) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] pdf and postscript sizes - change on each page?
Hi All The device commands pdf and postscript allow you to specify the width and height of a page. However, each subsequent plot is drawn on a separate page. Is there a way to change the page size part way through? For instance, is there an equivalent to the function pdfresize below? pdf(/tmp/1.pdf, width=6, height=6) plot(1:10) pdf.resize(width=8, height=5) plot(sin, -pi, pi) dev.off() --- So far, the only way I can find to do this is: pdf(/tmp/1.pdf, width=6, height=6) plot(1:10) dev.off() pdf(/tmp/1.pdf, width=8, height=5) plot(sin, -pi, pi) dev.off() --- however, this creates a new file - it does not append to the existing file. I would note that the quartz device has some resize capability, when you drag the window frame. -Alex __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exact 95% confidence intervals
check package binom. 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/(0)16/336899 Fax: +32/(0)16/337015 Web: http://med.kuleuven.be/biostat/ http://www.student.kuleuven.be/~m0390867/dimitris.htm - Original Message - From: XinMeng [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Sent: Wednesday, September 27, 2006 10:12 AM Subject: [R] exact 95% confidence intervals Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Thanks a lot! My best! -- *** Xin Meng Capitalbio Corporation National Engineering Research Center for Beijing Biochip Technology BioPharma-informatics Software Dept. Research Engineer Tel: +86-10-80715888/80726868-6438 Fax: +86-10-80726790 [EMAIL PROTECTED] Address:18 Life Science Parkway, Changping District, Beijing 102206, China __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] equivalent of model.tables for an lm.object?
Henrik Parn wrote: Dear all, I run a linear model with three significant explanatory variabels x1: a factor with 4 levels x2 and x3: factors with two levels each x4: continuous model - lm(y ~ x1 + x2 * x3 + x4) The data is not perfectly balanced between the different factor-combinations and I use treatment contrasts. With an aov.object, I assume I could have used model.tables(aov.object, type = means, se = TRUE), to get the means and se for all factor combinations. In an lm.object like mine, I calculate the means 'manually' from the Estimates (for sure it could be done with a script, but fair enough). For the standard error of the means, I started out using formulas of a variance of a sum of two variables, but I messed things up with the interaction. Is there a way to calculate the standard error of the means from Estimates and Std.Error (or other information) from the lm.object? You might have a look at the effects package by John Fox. http://socserv.mcmaster.ca/jfox/Misc/effects/index.html Thanks in advance for any advice! Best regards, Henrik -- Chuck Cleland, Ph.D. NDRI, Inc. 71 West 23rd Street, 8th floor New York, NY 10010 tel: (212) 845-4495 (Tu, Th) tel: (732) 512-0171 (M, W, F) fax: (917) 438-0894 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Searching for keyword values in a text (configuration) file
Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Searching for keyword values in a text (configuration) file
Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exponential fitting
# using this test data set.seed(1) x - 1:20/20 y - exp(2 + 3 * x) + rnorm(20) # if its ok to fit logs so that its linear exp(fitted(lm(log(y) ~ x))) 1 2 3 4 5 6 7 8 8.55615 9.94692 11.56376 13.44340 15.62857 18.16894 21.12223 24.7 910111213141516 28.54699 33.18720 38.58165 44.85295 52.14363 60.61938 70.47284 81.92793 17181920 95.24501 110.72673 128.72494 149.64869 # or to do it on original scale use linear coefs as starting values cc - coef(lm(log(y) ~ x)) fitted(nls(y ~ exp(a + b*x), start = list(a = cc[1], b = cc[2]))) [1] 8.592270 9.984536 11.602401 13.482421 15.667073 18.205720 [7] 21.155722 24.583734 28.567211 33.196159 38.575168 44.825776 [13] 52.089214 60.529599 70.337640 81.734946 94.979039 110.369167 [19] 128.253066 149.034820 attr(,label) [1] Fitted values On 9/27/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I would like to fit some experimental points by a exponential function. I ignore the parameters of this exponential and what I would like is to ask R to calculate the best fitting curve an the associated parameters (as the linear model function (lm) does for linear models). Is it possible ? Do anyone have an idea about how to do that ? Thanks by advance Jessica Gervais [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exact 95% confidence intervals
Frank E Harrell Jr [EMAIL PROTECTED] writes: Peter Dalgaard wrote: XinMeng [EMAIL PROTECTED] writes: Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Just run binom.test on the data from each column (or row, depending on which way you turn the tables). But beware of exact binomial intervals - they are often not very accurate. Wilson and other intervals are generally better. For example see the binconf function in the Hmisc package. I suppose that by accurate you mean that they are generally better at getting the coverage rate right? The exact intervals are strictly conservative, but at least predictably so. The whole thing is largely a matter of taste to my mind, but I know that other people (notably Alan Agresti) have stronger opinions. (People taking an interest in this may want to have a look at http://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval ) -- O__ Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
Re: [R] Impossible to merge with a zero rows data frame?
Hi you have two options: change the source code for merge or use some modified function like my.merge - function(x,y, ...) if(all(dim(x)[1]0, dim(y)[1]0)) merge(x,y) else print (No merge or whatever action which is suitable) HTH Petr On 27 Sep 2006 at 11:55, Bonfigli Sandro wrote: Date sent: Wed, 27 Sep 2006 11:55:44 +0200 From: Bonfigli Sandro [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Subject:[R] Impossible to merge with a zero rows data frame? I'm trying to merge two data frames. One of them is a zero rows data frame. I'm using the merge parameter 'all.x = TRUE' so I'd expect to obtain all the rows of x. In fact the merge help says: all.x: logical; if 'TRUE', then extra rows will be added to the output, one for each row in 'x' that has no matching row in 'y'. These rows will have 'NA's in those columns that are usually filled with values from 'y'. To let you test the problem here is some code: L3 - LETTERS[1:3] (d - data.frame(cbind(x=1, y=1), fac=sample(L3, 1, repl=TRUE))) e - d[-1,] so now we have that: e [1] x y fac 0 rows (or 0-length row.names) here is the merge: merge(d, e, by.x = c(x), by.y = c(x), all.x = TRUE) I'd expect something like: x y.x fac.x y.y fac.y 1 1 1 B NANA instead of: Error in merge.data.frame(d, e, by.x = c(x), by.y = c(x), all.x = TRUE) : no row to correspond (I'm traslating the error message, so it could be a little different) My questions are: Is there a way to obtain my desired result? In the context in which I'd like to use the code the second data frame is the result of a query and a lot of overwork would be added if I have to check if the dataframe is a zero rows one BEFORE the merge (in fact I do a sequence of several merge) Is this behaviour of the merge command correct? Why did the developers choose it? P.S.: I tested this code both in R 2.3.0 and in 2.3.1 Thank you in advance. Sandro Bonfigli __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Petr Pikal [EMAIL PROTECTED] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Searching for keyword values in a text (configuration) file
Read in data using readLines and replace = with comma and delete all spaces. Then reread using read.table and set the rownames to column 1 removing column 1. # test data Lines0 - DEVICE = 'PC' CPU_SPEED = '1999', '233' # if reading from a file then # replace next line with something like Lines - readLines(myfile.dat) Lines - readLines(textConnection(Lines0)) Lines - gsub(=, ,, Lines) Lines - gsub( , , Lines) DF - read.table(textConnection(Lines), sep = ,, fill = TRUE, colClasses = character, header = FALSE) rownames(DF) - DF[,1] DF - DF[,-1] DF[CPU_SPEED, 2] On 9/27/06, Andre Jung [EMAIL PROTECTED] wrote: Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Testing the equality of correlations
Dear All, I wonder if there is any implemented statistical test in R to test the equality between many correlations. As an example, let X1, X2, X3 X4 be four random variables. let Phi(X1,X2) , Phi(X1,X3) and Phi(X1,X4) be the corresponding correlations. How to test Phi(X1,X2) = Phi(X1,X3) = P(X1,X4)? Many thanks in advance, Bernard - [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Searching for keyword values in a text (configuration) file
Something like this? library(Hmisc) t - readLines(clipboard) t [1] DEVICE = 'PC' CPU_SPEED = '1999', '233' ix - grep(CPU_SPEED,t) loc - substring.location(t[ix],,) cpu - substring(t[ix],loc$first+2) cpu [1] '233' On 27/09/06, Andre Jung [EMAIL PROTECTED] wrote: Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- = David Barron Said Business School University of Oxford Park End Street Oxford OX1 1HP __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Searching for keyword values in a text (configuration) file
Here is one more solution using the same Lines0 from last time. This one uses strapply from gsubfn to pick out all the fields in each line creating a list named by the keywords (rather than a data frame as in the previous solution). The names of the components are the keywords so we remove them from the contents of the list itself since we don't need them twice: library(gsubfn) # replace next line with something like Lines - readLines(myfile.dat) Lines - readLines(textConnection(Lines0)) parms - strapply(Lines, [^ ',=]+, c, USE.NAMES = TRUE) parms - lapply(parms, [, -1) parms[[CPU_SPEED]][2] On 9/27/06, Gabor Grothendieck [EMAIL PROTECTED] wrote: Read in data using readLines and replace = with comma and delete all spaces. Then reread using read.table and set the rownames to column 1 removing column 1. # test data Lines0 - DEVICE = 'PC' CPU_SPEED = '1999', '233' # if reading from a file then # replace next line with something like Lines - readLines(myfile.dat) Lines - readLines(textConnection(Lines0)) Lines - gsub(=, ,, Lines) Lines - gsub( , , Lines) DF - read.table(textConnection(Lines), sep = ,, fill = TRUE, colClasses = character, header = FALSE) rownames(DF) - DF[,1] DF - DF[,-1] DF[CPU_SPEED, 2] On 9/27/06, Andre Jung [EMAIL PROTECTED] wrote: Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] PDE
Dear all, Does any know how to solve PDE with R? The archive list refers to the use of ODE if PDE are parabolic. I am not a mathematician and this does not mean anything for me! help would be very appreciated. Many thanks -- ___ Christophe NGUYEN UMR 1220 INRA-ENITAB Transfert sol-plante et cycle des éléments minéraux dans les écosystèmes cultivés Centre INRA de Bordeaux-Aquitaine 71, avenue Edouard Bourlaux, BP 81 33883 Villenave d'Ornon, FRANCE Tel : 00 33 (0)5 57 12 25 07 Fax : 00 33 (0)5 57 12 25 15 email : [EMAIL PROTECTED] page infoservice: http://www.bordeaux.inra.fr/tcem __m°O°m __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Searching for keyword values in a text (configuration) file
Here is a slight simplification of this one using row.names= in the read.table to avoid the subsequent manipulations. We use Lines0 defined in the earlier post: # replace next line with something like Lines - readLines(myfile.dat) Lines - readLines(textConnection(Lines0)) Lines - gsub(=, ,, Lines) Lines - gsub( , , Lines) DF - read.table(textConnection(Lines), row.names = 1, sep = ,, fill = TRUE, colClasses = character, header = FALSE) DF[CPU_SPEED, 2] On 9/27/06, Gabor Grothendieck [EMAIL PROTECTED] wrote: Read in data using readLines and replace = with comma and delete all spaces. Then reread using read.table and set the rownames to column 1 removing column 1. # test data Lines0 - DEVICE = 'PC' CPU_SPEED = '1999', '233' # if reading from a file then # replace next line with something like Lines - readLines(myfile.dat) Lines - readLines(textConnection(Lines0)) Lines - gsub(=, ,, Lines) Lines - gsub( , , Lines) DF - read.table(textConnection(Lines), sep = ,, fill = TRUE, colClasses = character, header = FALSE) rownames(DF) - DF[,1] DF - DF[,-1] DF[CPU_SPEED, 2] On 9/27/06, Andre Jung [EMAIL PROTECTED] wrote: Hi, I would like to read values from an ASCII text file that contains information in the following format: DEVICE = 'PC' CPU_SPEED = '1999', '233' ... It's like a config file. How can I e.g. get R to read the 2nd value of CPU_SPEED? How do I go through text files and search for keywords and their values? Thanks a lot, Andre __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] t-stat Curve
Number of subjects = 25 Mean of Sample = 77 Standard Deviation (s) = 12 sem = 2.4 df = 24 The claim is that population mean is less than 80 * 80 So our H0 (null hupotheis) is * 80 qt(.95,24) [1] 1.710882 qt(0.05, 24) [1] -1.710882 tstat = -1.25 on t24 falls between 1.711 (.95,24) and *1.711 (.005,24) How Could I sketch t curve for the above data where my * would be at the center? Best Regards Isaac Dr. I. Barjis Assistant Professor Summer and Evening Coordinator Department of Biological Sciences Room P313 300 Jay Street Brooklyn, NY 11201 Phone: (718)2605285 Fax: (718)2548680 Fax: (718) 254-8595 Department Office http://websupport1.citytech.cuny.edu/Faculty/ibarjis BEGIN:VCARD VERSION:2.1 X-GWTYPE:USER FN:Isaac Barjis ORG:;Physical Biological Sciences EMAIL;WORK;PREF;NGW:[EMAIL PROTECTED] N:Barjis;Isaac TITLE:Assistant Professor END:VCARD BEGIN:VCARD VERSION:2.1 X-GWTYPE:USER FN:Isaac Barjis ORG:;Physical Biological Sciences EMAIL;WORK;PREF;NGW:[EMAIL PROTECTED] N:Barjis;Isaac TITLE:Assistant Professor END:VCARD __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Testing the equality of correlations
Marc, Off the top my head (i.e. this could all be horribly wrong), I think Anderson gave an asymptotic version for such a test, whereby under the null hypothesis, the difference between Fisher's z for each sample, z1 - z2, is normal with zero mean. If I recall correctly, the 1984 edition gave a test statistic something like: $\frac{\vert z_{1} - z_{2} \rvert}{ \sqrt{\frac{1}{N_{1} - 3} + \frac{1}{N_{2} - 3}}}$, as the test. I ***think*** N = n+1. Assuming that part is correct, it can be coded up quite simply (and crudely and hopefully correctly) as: ## estimate Fisher's z for the sample fishz - function(x1,x2){ r - cor(x1,x2) z - 0.5 * log( (1+r)/(1-r) ) } ## apply the correlation test cortest - function(x1,x2,x3,x4){ numer - abs(fishz(x1,x2) - fishz(x3,x4) ) denom - sqrt( 1/( length(x1) - 2) + 1/( length(x2) - 2) ) test.stat - numer / denom return(test.stat) } A quick demo with some simulated data: require(MASS) X1 - mvrnorm(1000, c(0,0), matrix(c(1,0.7,0.7,1),2,2)) X2 - mvrnorm(1000, c(0,0), matrix(c(1,0.9,0.9,1),2,2)) cortest(X1[,1], X1[,2], X2[,1], X2[,2]) Is above 1.96 indicating they are different X2 - mvrnorm(1000, c(0,0), matrix(c(1,0.7,0.7,1),2,2)) cortest(X1[,1], X1[,2], X2[,1], X2[,2]) Is below 1.96 indicating they are not different. All that needs to be done for these pair is to get the pooled estimate of the populations Fisher's z $ \frac{ (N_{1} - 3)z_{1} + (N_{2} - 3)z_{2}}{N_{1} + N_{2} - 6}$ and solve fisher's z to get an estimate of rho. Just dabbling around with this suggests either I've missed something, or that we need quite a large sample size before the asymptotics are any use. If this is of any use I'll double check on N. Paul -=-=-=-=-=-=-=-=-=-=-=-= Paul Hewson Lecturer in Statistics School of Mathematics and Statistics University of Plymouth Drake Circus Plymouth PL4 8AA tel (01752) 232778 (Campus) tel (01752) 764437 (Tamar Science Park) fax (01752) 232780 email: [EMAIL PROTECTED] web: http://www.plymouth.ac.uk/staff/phewson -=-=-=-=-=-=-=-=-=-=-=-= -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Marc Bernard Sent: 27 September 2006 14:42 To: r-help@stat.math.ethz.ch Subject: [R] Testing the equality of correlations Dear All, I wonder if there is any implemented statistical test in R to test the equality between many correlations. As an example, let X1, X2, X3 X4 be four random variables. let Phi(X1,X2) , Phi(X1,X3) and Phi(X1,X4) be the corresponding correlations. How to test Phi(X1,X2) = Phi(X1,X3) = P(X1,X4)? Many thanks in advance, Bernard - [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] multidimensional lists
In the process of moving a number of my scripts from MATLAB - R, I've discovered that there is no 'pure' equivalent of MATLAB's cell arrays, which I use quite often. Basically, I create matrices (as a cell array) where each element of the matrix is itself a matrix (e.g., 2x2 cell array where each element of the array is another matrix). I pass these cell arrays to various functions which then do clever things with the various matrices (of course) - basically, I need to be able to pass collections of matrices to functions to do various things, and I need to be able to control the dimensionality of the cell array to preserve some structural relationships among the matrices in the array. The cell array in MATLAB handles this with aplomb. So far, in R, I've used lists. Given (say) 4 matrices (A,B,C,D), in MATLAB I can use test={A,B,C,D} for a row vector cell array, or test={A;B;C;D} for a column vector cell array. In R, I get more or less the same thing using test=list(A,B,C,D) but this only gives me a row list. For a bunch of technical reasons, I need to be able to control the orientation (as noted)- this is especially true for n-dimensional cell arrays. In MATLAB, for example, I could generate a (say) 2x2 cell array using test={A B;C D} The only way I can figure out how to do this in R is using something like test=list(A,B,C,D); dim(test) c(2,2); This seems to work, but defaults to bycolumn (in other words, instead of A B C D I get A C B D ) So, I follow with test=t(test) as needed to flip the thing around to byrow. OK, so the question is - is there a better way? This *seems* to work, but I'm discovering that R is a lot like working with LaTeX (something I know much more about) - you can do most things, but there is often a more elegant way if you can figure out how to find out about it. Thanks in advance... __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] t-stat Curve
-Original Message- From: [EMAIL PROTECTED] [mailto:r-help- [EMAIL PROTECTED] On Behalf Of Isaac Barjis Sent: Wednesday, September 27, 2006 8:08 AM To: R-help@stat.math.ethz.ch Subject: [R] t-stat Curve Number of subjects = 25 Mean of Sample = 77 Standard Deviation (s) = 12 sem = 2.4 df = 24 The claim is that population mean is less than 80 * 80 So our H0 (null hupotheis) is * 80 qt(.95,24) [1] 1.710882 qt(0.05, 24) [1] -1.710882 tstat = -1.25 on t24 falls between 1.711 (.95,24) and *1.711 (.005,24) How Could I sketch t curve for the above data where my * would be at the center? Best Regards Isaac Dr. I. Barjis Assistant Professor Summer and Evening Coordinator Department of Biological Sciences Isaac, I'm not sure that what you are asking for is reasonable (or possible). It is the sampling distribution of your t-statistic that is distributed as t under the null hypothesis, not your observed data. Could you clarify what it is you wish to do? Dan Daniel J. Nordlund Research and Data Analysis Washington State Department of Social and Health Services Olympia, WA 98504-5204 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] package e1071 - class probabilities
Hello, I use the package e1071 for svm classification. I would like to get probabilities associated with predictions. But when I change the argument 'probability' (TRUE or FALSE) in functions svm.model() and predict() , I don't have the same prediction. Do you know why there are these differences? Thank you in advance for your help. Best regards. Vincent. - [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
matrix(some_list, nr, nc, byrow=TRUE) may be what you are looking for. R arrays can be of any vector type, including list. I'd get used to R's Fortran ordering rather than force transposes all the time. On Wed, 27 Sep 2006, Evan Cooch wrote: In the process of moving a number of my scripts from MATLAB - R, I've discovered that there is no 'pure' equivalent of MATLAB's cell arrays, which I use quite often. Basically, I create matrices (as a cell array) where each element of the matrix is itself a matrix (e.g., 2x2 cell array where each element of the array is another matrix). I pass these cell arrays to various functions which then do clever things with the various matrices (of course) - basically, I need to be able to pass collections of matrices to functions to do various things, and I need to be able to control the dimensionality of the cell array to preserve some structural relationships among the matrices in the array. The cell array in MATLAB handles this with aplomb. So far, in R, I've used lists. Given (say) 4 matrices (A,B,C,D), in MATLAB I can use test={A,B,C,D} for a row vector cell array, or test={A;B;C;D} for a column vector cell array. In R, I get more or less the same thing using test=list(A,B,C,D) but this only gives me a row list. For a bunch of technical reasons, I need to be able to control the orientation (as noted)- this is especially true for n-dimensional cell arrays. In MATLAB, for example, I could generate a (say) 2x2 cell array using test={A B;C D} The only way I can figure out how to do this in R is using something like test=list(A,B,C,D); dim(test) c(2,2); This seems to work, but defaults to bycolumn (in other words, instead of A B C D I get A C B D ) So, I follow with test=t(test) as needed to flip the thing around to byrow. OK, so the question is - is there a better way? This *seems* to work, but I'm discovering that R is a lot like working with LaTeX (something I know much more about) - you can do most things, but there is often a more elegant way if you can figure out how to find out about it. Thanks in advance... __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
Try this: AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] On 9/27/06, Evan Cooch [EMAIL PROTECTED] wrote: In the process of moving a number of my scripts from MATLAB - R, I've discovered that there is no 'pure' equivalent of MATLAB's cell arrays, which I use quite often. Basically, I create matrices (as a cell array) where each element of the matrix is itself a matrix (e.g., 2x2 cell array where each element of the array is another matrix). I pass these cell arrays to various functions which then do clever things with the various matrices (of course) - basically, I need to be able to pass collections of matrices to functions to do various things, and I need to be able to control the dimensionality of the cell array to preserve some structural relationships among the matrices in the array. The cell array in MATLAB handles this with aplomb. So far, in R, I've used lists. Given (say) 4 matrices (A,B,C,D), in MATLAB I can use test={A,B,C,D} for a row vector cell array, or test={A;B;C;D} for a column vector cell array. In R, I get more or less the same thing using test=list(A,B,C,D) but this only gives me a row list. For a bunch of technical reasons, I need to be able to control the orientation (as noted)- this is especially true for n-dimensional cell arrays. In MATLAB, for example, I could generate a (say) 2x2 cell array using test={A B;C D} The only way I can figure out how to do this in R is using something like test=list(A,B,C,D); dim(test) c(2,2); This seems to work, but defaults to bycolumn (in other words, instead of A B C D I get A C B D ) So, I follow with test=t(test) as needed to flip the thing around to byrow. OK, so the question is - is there a better way? This *seems* to work, but I'm discovering that R is a lot like working with LaTeX (something I know much more about) - you can do most things, but there is often a more elegant way if you can figure out how to find out about it. Thanks in advance... __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] panel.curve
Hi, I am trying to fit experimental points by exponemtial curve my data are stored into a matrix data the first column is the geographical point (a number = data[,1] ) ( I would like to plot several graphes at the same time) the second column is the time of measurement (x in the plot) the third column is a speed (y in the plot) if we assume the point are folowing this exponential behaviour y=exp(a+bx) then log y = a+ bx we then can determine the coefficient a and b by a linear regression with the lm function and get them as following : coef ( lm (log(y)~x)) then I can use those coefficient if I plot ln y = ax+b , everything goes fine xyplot(log(data[,3])~data[,2]|data[,1],panel=function(x,y){panel.xyplot(x,y)+panel.abline(coef(lm(y~x)))}) and I get perfect linear regression of my points ...But I would prefer to plot the exponential curves (y=exp ( a*x + b )).. I tried the following formula : xyplot(data[,3]~data[,2]|data[,1],panel=function(x,y){panel.xyplot(x,y)+panel.curve(coef(lm(log(y)~x))[1])}) and I get : Error in model.frame(formula, rownames, variables, varnames, extras, extranames, : variable lengths differ (found for 'x') ... I don't really now what goes wrong and how to correct that Maybe I am wrong in the use of the pannel.curve function Do anyone know something about that ? Thanks by advance Jessica Gervais [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] t-stat Curve
Isaac: You will likely find something helpful here: http://addictedtor.free.fr/graphiques/thumbs.php I also recently came across this code (I thought it was at the URL above, but I can't find it now) that may be useful with modification. I apologize to the code-writer for having lost the correct reference. (If anyone finds it, please send the reference to me. Thanks.) # # neighboring (not overlapping) normal densities dev.off() x-seq(-10,10,length=400) y1-dnorm(x) y2-dnorm(x,m=3) par(mar=c(5,4,2,1)) plot(x, y2, xlim=c(-3,8), type=n, xlab=quote(Z==frac(mu[1]-mu[2], sigma/sqrt(n))), ylab=Density) polygon(c(1.96,1.96,x[240:400],10), c(0,dnorm(1.96,m=3),y2[240:400],0), col=grey80, lty=0) lines(x, y2) lines(x, y1) polygon(c(-1.96,-1.96,x[161:1],-10), c(0,dnorm(-1.96,m=0), y1[161:1],0), col=grey30, lty=0) polygon(c(1.96, 1.96, x[240:400], 10), c(0,dnorm(1.96,m=0), y1[240:400],0), col=grey30) legend(4.2, .4, fill=c(grey80,grey30), legend=expression(P(abs(phantom(i)*Z*phantom(i))1.96, H[1])==0.85, P(abs(phantom(i)*Z*phantom(i))1.96,H[0])==0.05), bty=n) text(0, .2, quote(H[0]:~~mu[1]==mu[2])) text(3, .2, quote(H[1]:~~mu[1]==mu[2]+delta)) # Charles Annis, P.E. [EMAIL PROTECTED] phone: 561-352-9699 eFax: 614-455-3265 http://www.StatisticalEngineering.com -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Isaac Barjis Sent: Wednesday, September 27, 2006 11:08 AM To: R-help@stat.math.ethz.ch Subject: [R] t-stat Curve Number of subjects = 25 Mean of Sample = 77 Standard Deviation (s) = 12 sem = 2.4 df = 24 The claim is that population mean is less than 80 * 80 So our H0 (null hupotheis) is * 80 qt(.95,24) [1] 1.710882 qt(0.05, 24) [1] -1.710882 tstat = -1.25 on t24 falls between 1.711 (.95,24) and *1.711 (.005,24) How Could I sketch t curve for the above data where my * would be at the center? Best Regards Isaac Dr. I. Barjis Assistant Professor Summer and Evening Coordinator Department of Biological Sciences Room P313 300 Jay Street Brooklyn, NY 11201 Phone: (718)2605285 Fax: (718)2548680 Fax: (718) 254-8595 Department Office http://websupport1.citytech.cuny.edu/Faculty/ibarjis __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] panel.curve
Look at the arguments to panel.curve ?panel.curve and try RSiteSearch(panel.curve) for some examples. On 9/27/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I am trying to fit experimental points by exponemtial curve my data are stored into a matrix data the first column is the geographical point (a number = data[,1] ) ( I would like to plot several graphes at the same time) the second column is the time of measurement (x in the plot) the third column is a speed (y in the plot) if we assume the point are folowing this exponential behaviour y=exp(a+bx) then log y = a+ bx we then can determine the coefficient a and b by a linear regression with the lm function and get them as following : coef ( lm (log(y)~x)) then I can use those coefficient if I plot ln y = ax+b , everything goes fine xyplot(log(data[,3])~data[,2]|data[,1],panel=function(x,y){panel.xyplot(x,y)+panel.abline(coef(lm(y~x)))}) and I get perfect linear regression of my points ...But I would prefer to plot the exponential curves (y=exp ( a*x + b )).. I tried the following formula : xyplot(data[,3]~data[,2]|data[,1],panel=function(x,y){panel.xyplot(x,y)+panel.curve(coef(lm(log(y)~x))[1])}) and I get : Error in model.frame(formula, rownames, variables, varnames, extras, extranames, : variable lengths differ (found for 'x') ... I don't really now what goes wrong and how to correct that Maybe I am wrong in the use of the pannel.curve function Do anyone know something about that ? Thanks by advance Jessica Gervais [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
Prof Brian Ripley wrote: matrix(some_list, nr, nc, byrow=TRUE) may be what you are looking for. R arrays can be of any vector type, including list. I'd get used to R's Fortran ordering rather than force transposes all the time. Thanks very much. And I thought I'd left some aspects of my FORTRAN coding experience in the past. Would seem not. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
Gabor Grothendieck wrote: Try this: AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] Seems to do the trick. Thanks! __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] exact 95% confidence intervals
Peter Dalgaard wrote: Frank E Harrell Jr [EMAIL PROTECTED] writes: Peter Dalgaard wrote: XinMeng [EMAIL PROTECTED] writes: Hello sir: As to the 2*2 table format for reporting results comparing a new test to true diagnosis,when I got the sensitivity and specificity,how can I calculate the exact 95% confidence intervals (based on the binomial distribution) for sensitivity and specificity via R? Just run binom.test on the data from each column (or row, depending on which way you turn the tables). But beware of exact binomial intervals - they are often not very accurate. Wilson and other intervals are generally better. For example see the binconf function in the Hmisc package. I suppose that by accurate you mean that they are generally better at getting the coverage rate right? Yes The exact intervals are strictly conservative, but at least predictably so. The whole thing is largely a matter of taste to my mind, but I know that other people (notably Alan Agresti) have stronger opinions. I tend to side with Agresti, and emphasize the expected absolute error in confidence coverage. (People taking an interest in this may want to have a look at http://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval ) nice to know! Frank __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
Just one other comment if the matrices have the same dimensions: they could alternately be represented as a 4d array: A - matrix(1:4, 2) AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] AAA - array(unlist(AA), c(2,2,2,2)) AAA[,,1,2] # same This could have an advantage if you need to do things like easily take the top left element of each matrix, AAA[1,1,,] which in terms of AA would have required the longer expression: replace(AA, TRUE, lapply(AA, [, 1, 1)) On 9/27/06, Gabor Grothendieck [EMAIL PROTECTED] wrote: Try this: AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] On 9/27/06, Evan Cooch [EMAIL PROTECTED] wrote: In the process of moving a number of my scripts from MATLAB - R, I've discovered that there is no 'pure' equivalent of MATLAB's cell arrays, which I use quite often. Basically, I create matrices (as a cell array) where each element of the matrix is itself a matrix (e.g., 2x2 cell array where each element of the array is another matrix). I pass these cell arrays to various functions which then do clever things with the various matrices (of course) - basically, I need to be able to pass collections of matrices to functions to do various things, and I need to be able to control the dimensionality of the cell array to preserve some structural relationships among the matrices in the array. The cell array in MATLAB handles this with aplomb. So far, in R, I've used lists. Given (say) 4 matrices (A,B,C,D), in MATLAB I can use test={A,B,C,D} for a row vector cell array, or test={A;B;C;D} for a column vector cell array. In R, I get more or less the same thing using test=list(A,B,C,D) but this only gives me a row list. For a bunch of technical reasons, I need to be able to control the orientation (as noted)- this is especially true for n-dimensional cell arrays. In MATLAB, for example, I could generate a (say) 2x2 cell array using test={A B;C D} The only way I can figure out how to do this in R is using something like test=list(A,B,C,D); dim(test) c(2,2); This seems to work, but defaults to bycolumn (in other words, instead of A B C D I get A C B D ) So, I follow with test=t(test) as needed to flip the thing around to byrow. OK, so the question is - is there a better way? This *seems* to work, but I'm discovering that R is a lot like working with LaTeX (something I know much more about) - you can do most things, but there is often a more elegant way if you can figure out how to find out about it. Thanks in advance... __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multidimensional lists
Gabor Grothendieck wrote: Just one other comment if the matrices have the same dimensions: Indeed - that is quite often the case - thanks for the further suggestion(s). they could alternately be represented as a 4d array: A - matrix(1:4, 2) AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] AAA - array(unlist(AA), c(2,2,2,2)) AAA[,,1,2] # same This could have an advantage if you need to do things like easily take the top left element of each matrix, AAA[1,1,,] which in terms of AA would have required the longer expression: replace(AA, TRUE, lapply(AA, [, 1, 1)) On 9/27/06, Gabor Grothendieck [EMAIL PROTECTED] wrote: Try this: AA - matrix(list(A, 10*A, 100*A, 1000*A), 2, byrow = TRUE) AA[1,2] On 9/27/06, Evan Cooch [EMAIL PROTECTED] wrote: In the process of moving a number of my scripts from MATLAB - R, I've discovered that there is no 'pure' equivalent of MATLAB's cell arrays, which I use quite often. Basically, I create matrices (as a cell array) where each element of the matrix is itself a matrix (e.g., 2x2 cell array where each element of the array is another matrix). I pass these cell arrays to various functions which then do clever things with the various matrices (of course) - basically, I need to be able to pass collections of matrices to functions to do various things, and I need to be able to control the dimensionality of the cell array to preserve some structural relationships among the matrices in the array. The cell array in MATLAB handles this with aplomb. So far, in R, I've used lists. Given (say) 4 matrices (A,B,C,D), in MATLAB I can use test={A,B,C,D} for a row vector cell array, or test={A;B;C;D} for a column vector cell array. In R, I get more or less the same thing using test=list(A,B,C,D) but this only gives me a row list. For a bunch of technical reasons, I need to be able to control the orientation (as noted)- this is especially true for n-dimensional cell arrays. In MATLAB, for example, I could generate a (say) 2x2 cell array using test={A B;C D} The only way I can figure out how to do this in R is using something like test=list(A,B,C,D); dim(test) c(2,2); This seems to work, but defaults to bycolumn (in other words, instead of A B C D I get A C B D ) So, I follow with test=t(test) as needed to flip the thing around to byrow. OK, so the question is - is there a better way? This *seems* to work, but I'm discovering that R is a lot like working with LaTeX (something I know much more about) - you can do most things, but there is often a more elegant way if you can figure out how to find out about it. Thanks in advance... __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] AIC Methods
I am having trouble with getting AIC statistics. I have developed a number of ancova models using the lm function. I am trying to get a number of AIC statistics suggested by Burnham and Anderson (2002) including raw AIC values, AICc values, Akaike weights, and Delta AIC. When I use the AIC function in (stats), I only get a single AIC value including degrees of freedom. When I use the summary statement as suggested in the (aod) package, I do not get the resulting AIC statistics. What am I doing wrong? Is this a product of using the lm function for my models? I have even investigated the use of extractAIC, but that does not provide me with the statistics I am looking for as well. Do you have any suggestions? In addition, I am also curious if there is a package or procedure for model averaging. Thank you! -- Benjamin Zuckerberg Doctoral Candidate State University of New York College of Environmental Science and Forestry Illick 244A, 1 Forestry Drive Syracuse, New York 13210 Tele: (315) 470-6985 E-mail: [EMAIL PROTECTED] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] PDE
Hi Christophe, What is the PDE that you are trying to solve? Is it parabolic/hyperbolic/elliptical/somethingelse? Is it linear/nonlinear? If time is one of the independent variables, you can transform the PDE into an initial value problem (system of ODEs) by using finite difference approximations of the partial derivatives of other independent variables (typically, these are spatial coordinates). Starting with an initial set of values on a grid of points (also known as initial conditions, which are part of the problem specification), you update them at different times, using fixed or varying time steps. R has very limited functionality for handling differential equations. So, you should look for FORTRAN libraries, from which you can create DLLs to be used in R. Hope this help, Ravi. --- Ravi Varadhan, Ph.D. Assistant Professor, The Center on Aging and Health Division of Geriatric Medicine and Gerontology Johns Hopkins University Ph: (410) 502-2619 Fax: (410) 614-9625 Email: [EMAIL PROTECTED] Webpage: http://www.jhsph.edu/agingandhealth/People/Faculty/Varadhan.html -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Christophe Nguyen Sent: Wednesday, September 27, 2006 10:39 AM To: r-help@stat.math.ethz.ch Subject: [R] PDE Dear all, Does any know how to solve PDE with R? The archive list refers to the use of ODE if PDE are parabolic. I am not a mathematician and this does not mean anything for me! help would be very appreciated. Many thanks -- ___ Christophe NGUYEN UMR 1220 INRA-ENITAB Transfert sol-plante et cycle des éléments minéraux dans les écosystèmes cultivés Centre INRA de Bordeaux-Aquitaine 71, avenue Edouard Bourlaux, BP 81 33883 Villenave d'Ornon, FRANCE Tel : 00 33 (0)5 57 12 25 07 Fax : 00 33 (0)5 57 12 25 15 email : [EMAIL PROTECTED] page infoservice: http://www.bordeaux.inra.fr/tcem __m°O°m __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Single Precision (4 byte) floats with readBin
I would like to use readBin to read a binary data file. Most of the data is 4-byte floating point but, for some reason, only double precision appears to be offered. I tried fVariable=readBin(iFile,what=single()); and got 35.87879 which looks believable except that the correct value is 3.030303. I then tried fVariable=readBin(iFile,what=single(),4); and got [1] 3.83e+10 6.657199e+10 -5.592394e+29 -5.592397e+29 For the second call, there were two more single precision floats of value 3.030303 followed by two more with values 40.46 and 0.00 respectively. Is there any way around this problem other than to make the input data double (which I definitely do not want to do)? Many thanks in advance, Peter. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] how to retain time zone when doing c(POSIXct)
Hello, What is the best way to concatenate POSIXct objects keeping the time zone attribute in a program? For example: R xx - as.POSIXct(strptime(c(2006-09-26 12:00:00, 2006-09-26 13:00:00), + format=%Y-%m-%d %H:%M:%S), tz=GMT) R xx [1] 2006-09-26 12:00:00 GMT 2006-09-26 13:00:00 GMT but, as ?c.POSIXct explains: R c(xx, xx[1] - 60, xx[2] + 60) [1] 2006-09-26 07:00:00 CDT 2006-09-26 08:00:00 CDT [3] 2006-09-26 06:59:00 CDT 2006-09-26 08:01:00 CDT Is there something better/safer than simply setting the tzone attribute of the new object? Cheers, -- Seb __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Single Precision (4 byte) floats with readBin
Peter Lauren said the following on 9/27/2006 3:11 PM: I would like to use readBin to read a binary data file. Most of the data is 4-byte floating point but, for some reason, only double precision appears to be offered. I tried fVariable=readBin(iFile,what=single()); and got 35.87879 which looks believable except that the correct value is 3.030303. I then tried fVariable=readBin(iFile,what=single(),4); and got [1] 3.83e+10 6.657199e+10 -5.592394e+29 -5.592397e+29 For the second call, there were two more single precision floats of value 3.030303 followed by two more with values 40.46 and 0.00 respectively. Is there any way around this problem other than to make the input data double (which I definitely do not want to do)? Many thanks in advance, Peter. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Hi, Peter, I believe you can use readBin(file, double(), size = 4) Thanks, --sundar __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Single Precision (4 byte) floats with readBin
--- Sundar Dorai-Raj [EMAIL PROTECTED] wrote: Peter Lauren said the following on 9/27/2006 3:11 PM: I would like to use readBin to read a binary data file. Most of the data is 4-byte floating point but, for some reason, only double precision appears to be offered. I tried fVariable=readBin(iFile,what=single()); and got 35.87879 which looks believable except that the correct value is 3.030303. I then tried fVariable=readBin(iFile,what=single(),4); and got [1] 3.83e+10 6.657199e+10 -5.592394e+29 -5.592397e+29 For the second call, there were two more single precision floats of value 3.030303 followed by two more with values 40.46 and 0.00 respectively. Is there any way around this problem other than to make the input data double (which I definitely do not want to do)? Many thanks in advance, Peter. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Hi, Peter, I believe you can use readBin(file, double(), size = 4) It worked! Thanks very much. Peter. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Converting text to numbers
Hi, I have Forecast Class and Observed Class in a data matrix as below. Sample1 FCT OBS 1 1 5 2 2 4 3 3- 3+ 4 3 3 5 3+ 3- 6 4 2 7 5 1 I want to find the difference between Observed and Forecast Classes. How can I get this done? I tried to following to convert the 1 through 5 classes, to 1 through 7 for both OBS and FCT column. Sample1$OBS2 - Sample1$OBS levels(Sample1$OBS2) - sub('5',7,levels(Sample1$OBS2),fixed=TRUE) levels(Sample1$OBS2) - sub('4',6,levels(Sample1$OBS2),fixed=TRUE) levels(Sample1$OBS2) - sub('3+',5,levels(Sample1$OBS2), fixed=TRUE) levels(Sample1$OBS2) - sub('3',4,levels(Sample1$OBS2),fixed=TRUE) levels(Sample1$OBS2) - sub('4-',3,levels(Sample1$OBS2),fixed=TRUE) Sample1 FCT OBS FCT2 OBS2 1 1 5 1 7 2 2 426 3 3- 3+35 4 3 344 5 3+ 3-53 6 4 262 7 5 171 All looks good, but as I do the following, I encounter an error. Sample1$OBS2- Sample1$FCT2 [1] NA NA NA NA NA NA NA Warning message: - not meaningful for factors in: Ops.factor(Sample1$OBS2, Sample1$FCT2) Then, I tried to convert them to numbers using the following. Sample1$FCT2 - as.numeric(Sample1$FCT2) Sample1$OBS2 - as.numeric(Sample1$OBS2) Sample1 FCT OBS FCT2 OBS2 1 1 5 17 2 2 426 3 3- 3+45 4 3 333 5 3+ 3-54 6 4 262 7 5 171 Sample1$FCT2[3] and Sample1$FCT2[4] switched values. I think it has something to do with the following: Sample1$OBS [1] 5 4 3+ 3 3- 2 1 Levels: 1 2 3 3- 3+ 4 5 But, I don't know why and how to fix it. Any ideas? Thank you. Daniel Chan Meteorologist Georgia Forestry Commission P O Box 819 Macon, GA 31202 Tel: 478-751-3508 Fax: 478-751-3465 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] t-stat Curve
## There is some ambiguity in your example. ## You stated a one-sided hypothesis and calculated qt() values for both sides. ## I show both the one-sided and two-sided displays. library(HH) ## HH_1.5 is available from CRAN for R-2.3.1 ## ## HH_1.5 ignores the df.t argument and interprets the request as a ## normal distribution. ## HH_1.8 has been accepted for CRAN for R-2.4.0 and will be in the ## standard places on CRAN when R-2.4.0 is released. ## ## HH_1.8 uses the df.t argument and interprets the request as a ## t-distribution. old.par - par(oma=c(4,0,2,5), mar=c(7,7,4,2)+.1) crit.val.t - qt(c(.05,.95), 24) crit.val - crit.val.t*(12/sqrt(25)) + 80 observed.t - -1.25 observed.ybar - 77 norm.setup(mean=80, n=25, sd=12, df.t=24, xlim=c(70,90), main=two-sided alpha=.10) norm.curve(mean=80, n=25, sd=12, df.t=24, crit=crit.val) abline(v=observed.ybar) axis(side=3, at=observed.ybar, line=-.5) norm.setup(mean=80, n=25, sd=12, df.t=24, xlim=c(70,90), main=one-sided alpha=.05) norm.curve(mean=80, n=25, sd=12, df.t=24, crit=crit.val[1], shade=left) abline(v=observed.ybar) axis(side=3, at=observed.ybar, line=-.5) par(old.par) __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Space required by object?
Does R provide a function analogous to LS() or str() that reports the storage space, on disk or in memory, required by objects? Ben Fairbank __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Converting text to numbers
Then, I tried to convert them to numbers using the following. Sample1$FCT2 - as.numeric(Sample1$FCT2) Sample1$OBS2 - as.numeric(Sample1$OBS2) This is actually an FAQ. Do the following and it should be fine: Sample1$FCT2 - as.numeric(as.character(Sample1$FCT2)) Sample1$OBS2 - as.numeric(as.character(Sample1$OBS2)) -- = David Barron Said Business School University of Oxford Park End Street Oxford OX1 1HP __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Space required by object?
Here is a function I use: my.ls - function(pos=1, sorted=F){ .result - sapply(ls(pos=pos, all.names=TRUE), function(..x)object.size(eval(as.symbol(..x if (sorted){ .result - rev(sort(.result)) } .ls - as.data.frame(rbind(as.matrix(.result),**Total=sum(.result))) names(.ls) - Size .ls$Size - formatC(.ls$Size, big.mark=',', digits=0, format='f') .ls$Mode - c(unlist(lapply(rownames(.ls)[-nrow(.ls)], function(x)mode(eval(as.symbol(x), '---') .ls } On 9/27/06, Ben Fairbank [EMAIL PROTECTED] wrote: Does R provide a function analogous to LS() or str() that reports the storage space, on disk or in memory, required by objects? Ben Fairbank __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem you are trying to solve? __ R-help@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 and provide commented, minimal, self-contained, reproducible code.
Re: [R] Space required by object?
See ll() in the R.oo package. /Henrik On 9/27/06, Ben Fairbank [EMAIL PROTECTED] wrote: Does R provide a function analogous to LS() or str() that reports the storage space, on disk or in memory, required by objects? Ben Fairbank __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Histogram
Dear all, I want to design a histogram and I need to have the frequency at certain points. For example I have the following 2 columns: *X Y* 0.125 0.422 0.45 11 0.55 21 I want the chart to have 4 columns. First column is from 0.0-0.1 (on X) and frequency is 25. Next colum is wider and form 0.1-0.4 with 22 frequency. Next column is narrow with 11 frequency and the last column is the same as the first one with 21 frequency. Can anybody tell me how I can have this chart. Thanks, Mohsen [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Histogram
Hi, There may be an easier way but here is one way you can do it. # create vector that has Y[i] X[i]s new.data - rep(X,Y) hist(new.data, breaks=c(0,.1,.4,.6)) # or something like that look at what exactly breaks should be. Ritwik. On 9/27/06, Mohsen Jafarikia [EMAIL PROTECTED] wrote: Dear all, I want to design a histogram and I need to have the frequency at certain points. For example I have the following 2 columns: *X Y* 0.125 0.422 0.45 11 0.55 21 I want the chart to have 4 columns. First column is from 0.0-0.1 (on X) and frequency is 25. Next colum is wider and form 0.1-0.4 with 22 frequency. Next column is narrow with 11 frequency and the last column is the same as the first one with 21 frequency. Can anybody tell me how I can have this chart. Thanks, Mohsen [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Ritwik Sinha Graduate Student Epidemiology and Biostatistics Case Western Reserve University [EMAIL PROTECTED] | +12163682366 | http://darwin.cwru.edu/~rsinha __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Histogram
On Wed, 2006-09-27 at 18:29 -0400, Mohsen Jafarikia wrote: Dear all, I want to design a histogram and I need to have the frequency at certain points. For example I have the following 2 columns: *X Y* 0.125 0.422 0.45 11 0.55 21 I want the chart to have 4 columns. First column is from 0.0-0.1 (on X) and frequency is 25. Next colum is wider and form 0.1-0.4 with 22 frequency. Next column is narrow with 11 frequency and the last column is the same as the first one with 21 frequency. Can anybody tell me how I can have this chart. Thanks, Mohsen How about this: X - c(0, 0.1, 0.4, 0.45, 0.55) Y - c(25, 22, 11, 21) barplot(Y, space = 0, width = diff(X)) axis(1) See ?barplot and ?diff HTH, Marc Schwartz __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Best use of LaTeX listings package for pretty printing R code
I tried to use Frank's formatting improvements to compile Sweave- produced .tex, but it doesn't seem to make a difference. Advice? On Sep 25, 2006, at 9:31 AM, Frank E Harrell Jr wrote: This is what I have been using. Does anyone have a better way? In particular I would like to see letters in comment strings not stretched so much. Thanks -Frank \documentclass{article} \usepackage{listings,relsize} \lstloadlanguages{R} \newcommand{\lil}[1]{\lstinline|#1|} \begin{document} \lstset{language=R,basicstyle=\smaller,commentstyle=\rmfamily\smaller, showstringspaces=false,% xleftmargin=4ex,literate={-}{{$\leftarrow$}}1 {~}{{$\sim$}}1} \lstset{escapeinside={(*}{*)}} % for (*\ref{ }*) inside lstlistings (S code) \begin{lstlisting} a - b # this is a test line if(i==3) { # another line, for y^2 y - 3^3 z - 'this string' qqcat - y ~ pol(x,2) } else y - 4 \end{lstlisting} That was \lstinline|x - 22| \lil{q - 'cat'}. \end{document} _ Professor Michael Kubovy University of Virginia Department of Psychology USPS: P.O.Box 400400Charlottesville, VA 22904-4400 Parcels:Room 102Gilmer Hall McCormick RoadCharlottesville, VA 22903 Office:B011+1-434-982-4729 Lab:B019+1-434-982-4751 Fax:+1-434-982-4766 WWW:http://www.people.virginia.edu/~mk9y/ __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] recode problem - unexplained values
I am hoping for some advice regarding the difficulties I have been having recoding variables which are contained in a csv file. Table 1 (below) shows there are two types of blanks - as reported in the first two columns. I am using windows XP the latets version of R. When blanks cells are replaced with a value of n using syntax: affect [affect==] - n there are still 3 blank values (Table 2). When as.numeric is applied, this also causes problems because values of 2,3 4 are generated rather than just 1 2. TABLE 1 table(group,actions) actions group n y 1 100 2 0 3 2 30 1 1 0 3 24 0 0 0 TABLE 2 table(group,actions) actions group n y 1 0 2 100 3 2 0 1 31 0 3 0 0 24 0 Below is another example - for some reason there are 2 types of 'aobh' values. table(group, type) type group aobh aobh gbh m uw 1 104 1 0 0 0 20 0 15 0 17 30 0 0 24 0 Any assistance is much appreciated, Bob Green __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Nonlinear fitting - reparametrization help
Hi, I am trying to fit a function of the form: y = A0 + A1 * exp( -0.5* ( (X - Mu1) / Sigma1 )^2 ) - A2 * exp ( -0.5* ( (X-Mu2)/Sigma2 )^2 ) i.e. a mean term (A0) + a difference between two gaussians. The constraints are A1,A2 0, Sigma1,Sigma20, and usually Sigma2Sigma1. The plot looks like a Mexican Hat. I had trouble (poor fits) fitting this function to toy data in Matlab and now I am playing with R's nls and optim functions. I keep running into singular gradient errors with nls, even with very different starting values, and have not yet figured out how to interpret the trace results usefully. Can someone help ? Is there a correct parameterization for this problem ? I have appended some R code with sample data to fit. Thank you !!! Suresh; please cc [EMAIL PROTECTED] if possible. ~~~ x=seq(-10,10,length=1000) a0=0; ae=1; me=0; se=1; ai=0.5; mi=0; si=3; dogy - function(x,a0,ae,me,se,ai,mi,si){ y=a0+ae*exp(-0.5*(((x-me)/se)^2))-ai*exp(-0.5*(((x-mi)/si)^2)) y} y=dogy(x,a0,ae,me,se,ai,mi,si) erval=rnorm(length(y),sd=0.02) y=y+erval #plot(x,y+erval) #fit=nls(y~ae*exp((x/se)^2)-ai*exp((x/si)^2),start=c(ae=.8,se=1.1,ai=.2,si=1),trace=TRUE) #here I tried to unsuccesfully reduce the model by eliminating A0, Mu1, and Mu2 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Nonlinear fitting - reparametrization help
Similar question was recently asked. See this thread: http://comments.gmane.org/gmane.comp.lang.r.general/69592 On 9/28/06, Suresh Krishna [EMAIL PROTECTED] wrote: Hi, I am trying to fit a function of the form: y = A0 + A1 * exp( -0.5* ( (X - Mu1) / Sigma1 )^2 ) - A2 * exp ( -0.5* ( (X-Mu2)/Sigma2 )^2 ) i.e. a mean term (A0) + a difference between two gaussians. The constraints are A1,A2 0, Sigma1,Sigma20, and usually Sigma2Sigma1. The plot looks like a Mexican Hat. I had trouble (poor fits) fitting this function to toy data in Matlab and now I am playing with R's nls and optim functions. I keep running into singular gradient errors with nls, even with very different starting values, and have not yet figured out how to interpret the trace results usefully. Can someone help ? Is there a correct parameterization for this problem ? I have appended some R code with sample data to fit. Thank you !!! Suresh; please cc [EMAIL PROTECTED] if possible. ~~~ x=seq(-10,10,length=1000) a0=0; ae=1; me=0; se=1; ai=0.5; mi=0; si=3; dogy - function(x,a0,ae,me,se,ai,mi,si){ y=a0+ae*exp(-0.5*(((x-me)/se)^2))-ai*exp(-0.5*(((x-mi)/si)^2)) y} y=dogy(x,a0,ae,me,se,ai,mi,si) erval=rnorm(length(y),sd=0.02) y=y+erval #plot(x,y+erval) #fit=nls(y~ae*exp((x/se)^2)-ai*exp((x/si)^2),start=c(ae=.8,se=1.1,ai=.2,si=1),trace=TRUE) #here I tried to unsuccesfully reduce the model by eliminating A0, Mu1, and Mu2 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] recode problem - unexplained values
I can propose a strategy. This example shows that there are different types of blanks when you look at character data. as.character(c(, , ,)) Your test for found only one of them. Look at the data as read.csv produces it. That will probably give you some clues. mydata - read.csv(filename) mydata as.character(mydata) Rich __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.