Re: [R] Contrasts for MANOVA
On Fri, 19 Dec 2003, Andy Bunn wrote: A colleague asked me if R can do contrasts for MANOVA. SAS will using PROC GLM with CONTRAST but can R? I assured her that R can do everything but make coffee. If this is trivial then I apologize in advance. Depends what you mean: I don't believe that SAS command does what I understand by manova, at least as its main purpose. Both lm() and aov() handle multiple responses. aov() in R is really set up to consider terms, but its summary method can split terms (via its `split' argument). What summary.manova does which the other aov methods does not is to compute multivariate test statistics for each term. If your colleague wants both the splitting allowed in summary.aov and the multivariate test statistics of summary.manova then that is not currently supported. What I would do there is to is ensure each of the contrasts I wanted had a separate term in the model, and then use summary(manova(...)). As ever, an example of what you want would help, as I have been guessing rather. -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] Error running LDA
As it says, some of your variables appear not to vary, which means LDA is not appropriate. LDA assumes a common multivariate normal distribution for the variation of each group about the group centre, and that cannot generate data like yours. On Fri, 19 Dec 2003, Yun-Fang Juan wrote: oh sorry here is the error. Please advice. thanks, Yun-Fang lda1 - lda(retention ~ . , data=RetentionDF40[1:1,]); Error in lda.default(x, grouping, ...) : variable(s) 11 14 24 29 30 31 32 33 34 35 41 43 44 47 48 49 50 51 54 55 56 57 58 59 60 61 65 66 67 68 70 71 72 74 78 82 85 87 88 89 91 92 93 101 102 103 107 108 109 111 112 117 122 124 125 127 148 151 153 155 156 162 170 173 176 181 182 183 184 186 189 190 192 193 194 196 200 209 212 213 214 215 216 218 220 222 223 224 228 230 231 234 235 236 237 appear to be constant within groups - Original Message - From: Marc R. Feldesman [EMAIL PROTECTED] To: Yun-Fang Juan [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Friday, December 19, 2003 6:38 PM Subject: Re: [R] Error running LDA At 06:13 PM 12/19/2003, Yun-Fang Juan wrote: Hi, I try to run the linear discriminant analysis using the following command but got an error like the following. lda1 - lda(retention ~ . , data=RetentionDF40[1:1,]); What error did you get? Did you read the help file for lda? What version of R and what version of VR are you using? What platform? __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] draft of posting guide
Tony Plate [EMAIL PROTECTED] writes: h3Common posting mistakes/h3 Doing any of the following may result in you getting a response that you may find rude or insulting. (However, such a response may be justified in the eyes of some because you have wasted people's time or unjustly insulted people's work.) * Not doing your homework before posting a question. * Asking R-help to do your classroom homework for you (remember that many of the subscribers of R-help are university professors and can recognize homework questions with great speed and accuracy) * Claiming that something is a bug when in fact the software is working as intended and documented, just not in the way you first expected. * Claiming that some commonly used function is not behaving in what you think is a sensible manner (it's far more productive and polite to just ask why it behaves the way it does if you think it is odd -- but only after reading all the relevant documentation!) * Threatening not to use the software if you cannot get your question answered. Even when intended as a statement of fact, it tends to create negative attitudes. -- O__ Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
RE: [R] draft of posting guide
Thanks for your effort. Here are some comments. 1. The guidelines seem only to cover people _asking_ questions. What about those answering? 2. I think bug needs to be taken in the broader sense of a problem. Such problems are not limited to discrepancies between documentation and implementation. The real serious problems are design problems and may not represent such deviations at all. If the behavior is unexpected it may very well be a problem even if its not a bug in the narrow sense. 3. Rude answers are never warranted. 4. Its too long and there are too many rules. 5. The tone should be friendlier -- more helpful and inviting. Words like lazy should be eliminated. Although R is not a commercial enterprise and its users are not customers in the usual sense, its users should still be treated with respect and encouraged rather than admonished. A customer oriented attitude will be appreciated and encourage more participation. The purpose of this should not be so much to tell posters what they can do and not do but to help them elicit useful responses and to help responders provide userful answers. 6. Points should be numbered so they can be referred to. Each one should have an address so that one can refer to them individually like this: http://www.r-project.org/...whatever.../etiquette.html#3.1 7. I personally find the statistics questions interesting and would not like to see guidelines about not posting statistical problems. 8. Where documents are referred to, provide a URL. Where R is being referred to point that out. For example, help.search is an R command. This may be obvious to some but it might not be obvious to someone just coming to R. 9. It might be worthwhile to point out that the following google search works: whatever site:r-project.org The real problem with the lists is not the posters. The real problems are: - rude answers - insufficient discussion about the direction and design and, in general, higher level issues. If you compare this to ruby, python, perl and lua there are all sorts of interesting proposals and discussions on this on their lists with little counterpart on R's. Date: Fri, 19 Dec 2003 22:55:53 -0700 From: Tony Plate [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: [R] draft of posting guide Here is a first draft of a guide for posters to r-help and r-devel. Suggestions on how to improve any aspect of it are most welcome. Suggestions on ways to make it more concise are especially welcome. Comments on which parts you like or don't like are welcome. Rather than clutter up R-help with lots of small corrections etc, please send them to me, and I will try to incorporate and summarize appropriately. I've left out most HTML formatting (except for headings) to make it readable -- I'll insert that later (to conform with the style on the other pages at r-project.org.) I've placed ??? in places where feedback is specifically needed. h2Posting Guide: How to ask good questions that prompt useful answers/h2 * Remember that R is free software, constructed and maintained by volunteers. They have various reasons for contributing to R, but often have limited time. Remember that no one owes you anything, and if you rude or disrespectful in your questions, you are likely to either be ignored or have the sentiment returned to you. * The R mailing lists are primarily intended for questions and discussion about the R software. However, sometimes questions about statistical methodology are posted. [??? are these discouraged/tolerated/... ???] People are far less likely to respond to these types of questions than to questions about the R software. Depending on how much the post shows thought and background research, responses may merely (and sometimes brusquely) suggest that you should be seeking statistical consulting advice elsewhere. Sometimes, if the question is well-asked AND of interest to someone on the list, you may get an informative, up-to-date answer. * Don't expect R-help to teach you basic statistics. That's what statistics textbooks and statistics classes are for. * If you have a question about R that you want answered, don't be lazy. Do your homework first. If it is clear that you have done your homework, your are far more likely to get an informative response. The type of homework that needs to be done depends on the type of question -- details are in the following paragraphs. * Make it easy for people to answer your question: be clear and concise, remove unnecessary details, and, if they might be useful, provide brief examples. * A much more detailed (and highly recommended) essay on how to ask questions on mailing lists about software is Eric Raymond's How To Ask Questions The Smart Way http://www.catb.org/~esr/faqs/smart-questions.html. (Note that catb.org has no association with the R project and will not respond to questions concerning R.)
[R] sound library
I'm collaborating with and electronic musician to experiment on the production of music from number sequences. As I'm an R user I started playing around with the sound library and I found it very useful. However there are several things I do not understand (I'm not an expert in acustic nor audio signal treatment). The first thing I'd like to understand is: let s be a normalized sample and suppose it generates a waveform with non audible frequencies. Then if I play it trough the function play (I'm using an R-window version) I won't hear anything. Then I play: play(1*s) and I can hear something. What is happening to the audio sample? I saved as wave files both s and 1*s and then with my musician friend we looked at the waveforms we got. They are totally different from each other and we couldn't fidure out which transformation of s generated the 1*s waveform. Can anyone explain what is going on? Thanks in advance Giovanna Jona Lasinio __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] sound library
[EMAIL PROTECTED] wrote: I'm collaborating with and electronic musician to experiment on the production of music from number sequences. As I'm an R user I started playing around with the sound library and I found it very useful. However there are several things I do not understand (I'm not an expert in acustic nor audio signal treatment). The first thing I'd like to understand is: let s be a normalized sample and suppose it generates a waveform with non audible frequencies. Then if I play it trough the function play (I'm using an R-window version) I won't hear anything. Then I play: play(1*s) and I can hear something. What is happening to the audio sample? Looks like it is not checked whether the generated sound, i.e. a wave file, is valid. Using 16 bit resolution, you cannot have more than 2^16 integers, but 1*s contains invalid integers for the wave representation. I saved as wave files both s and 1*s and then with my musician friend we looked at the waveforms we got. They are totally different from each other and we couldn't fidure out which transformation of s generated the 1*s waveform. Since you are writing invalid values to the file, those are interpreted in the wrong way and you might get something different. Can anyone explain what is going on? Two more comments: sound as is does not pass Rcmd check on Windows (because Media Player is not available on the system I do compile on, etc.) and has some more deficiencies. It's only available as a Windows binary on CRAN because of my personal interest in the package. The author of the package, Matthias Heymann, has been informed, but he is not going to fix the bugs in the near future, unfortunately. I am planing to write a package for the analysis of music (including loading / playing / saving wave files) next year myself. Uwe Ligges Thanks in advance Giovanna Jona Lasinio __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] for loop over dataframe without indices
I think I've found a problem with the by approach. Compare: data(iris) by( iris, row.names(iris), function(x)x )[1:5,] to iris[1:5,] It seems by has reordered the rows. Date: Fri, 19 Dec 2003 21:31:50 -0500 (EST) From: Gabor Grothendieck [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: [R] for loop over dataframe without indices Thomas, Thanks for your response. Its is quite nifty. Pursuing your solutions, I think the objective should be to reproduce the output from t.data.frame defined as below (note that I posted a proposal to change t.data.frame to r-devel before I received your reply): t.data.frame - function( df ) { ll - NULL for( i in 1:nrow(df) ) ll - append( ll, list(df[i,]) ) ll } Using the first 3 rows from the iris data set as our data frame, run the following which shows that your by solution works provided we nullify out the attributes afterwards. The do.call solution does not appear to work, as required, since it turns the data frame into a matrix. data(iris) df - iris[1:3,] # Consider: id - function(x)x # t.data.frame solution zt - t(df) # by solution is good but it adds some junk attributes zby - by( df, row.names(df), id ) identical(zt,zby) # FALSE # nullifying these attributes seems to do it zby2 - zby attributes(zby2) - NULL identical(zt,zby2) # TRUE # do.call doesn't work right since it appears to turn the result into a matrix str( do.call(mapply, list(id,df) ) ) # note matrix output Here is the result of pasting the above into R 1.8.1 on Windows 2000: data(iris) df - iris[1:3,] # Consider: id - function(x)x # t.data.frame solution zt - t(df) # by solution is good but it adds some junk attributes zby - by( df, row.names(df), id ) identical(zt,zby) [1] FALSE # nullifying these attributes seems to do it zby2 - zby attributes(zby2) - NULL identical(zt,zby2) [1] TRUE # do.call doesn't work right since it appears to turn the result into a matrix str( do.call(mapply, list(id,df) ) ) num [1:3, 1:5] 5.1 4.9 4.7 3.5 3 3.2 1.4 1.4 1.3 0.2 ... - attr(*, dimnames)=List of 2 ..$ : NULL ..$ : NULL Based on your solution I think the proposal should be changed to: t.data.frame - function(df) { z - by( df, row.names(df), function(x)x ) attributes(z) - NULL z } --- Date: Fri, 19 Dec 2003 10:03:55 -0800 (PST) From: Thomas Lumley [EMAIL PROTECTED] To: Gabor Grothendieck [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: [R] for loop over dataframe without indices On Fri, 19 Dec 2003, Gabor Grothendieck wrote: What I now realize is that the thing that is oddly missing in R is that you can't do an apply over the rows of a dataframe (at least not without having it coerced to an array and the elements coerced to possibly different types). The documentation does point this out. Its not a bug but its an omission that seems deserving of being addressed. Since mapply() applies a function to each 'row' of a list of vectors, ou can achieve this effect with do.call(mapply, list(FUN,data.frame)) and also as a degenerate case of by(): by(data.frame, row.names(data.frame), FUN) These should probably be documented under apply() -thomas __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] C++: Appending Values onto an R-Vector.
[Jim Java] Hi folks. I posted this question a few days ago, but maybe it got lost because of the code I included with it. I'm having a problem using the SET_LENGTH() macro in an R extension I'm writing in C++. In a function within the extension I use SET_LENGTH() to resize R vectors so as to allow the concatenation of single values onto the vectors -- it's a push back function to append values onto the end of a vector. However, when I use this function to push back a large number of values one at a time, Rgui.exe (I'm working with R 1.8.1 in Windows XP) crashes from an Access Violation; if, however, I pre-allocate space (is the space actually pre-allocated?) for the vector (say with NEW_INTEGER(n) rather than NEW_INTEGER(0)) and insert values into the allocated slots, the code works fine. If you'd like to see some test code, I've already posted it here: https://www.stat.math.ethz.ch/pipermail/r-help/2003-December/041871.html Here's my question, then: Is SET_LENGTH() the appropriate way to create space for tacking values onto the end of an R-vector in C++, or should I be trying to tack them on in some other way? [Brian Ripley] You UNPROTECT before calling lengthgets, and that is I think part of your problem. You need to use REPROTECT... [Jim Java] This problem is solved, so I want to post a follow-up for reference in the archives. My thanks to Prof. Ripley for helping me out: the problem was indeed in using UNPROTECT before lengthgets() rather than using REPROTECT afterwards. My revised test code is listed below, also for reference. CPP Code SEXP R_SimplePushBackTest(SEXP args) { SEXP arg1, arg2, int_vect; PROTECT(arg1 = AS_INTEGER(CADR(args))); int n_reps = INTEGER_POINTER(arg1)[0]; PROTECT(arg2 = AS_LOGICAL(CADDR(args))); bool full_alloc = (LOGICAL_POINTER(arg2)[0] ? true : false); PROTECT_INDEX int_vect_pindex; if (full_alloc) PROTECT_WITH_INDEX(int_vect = NEW_INTEGER(n_reps), int_vect_pindex); else PROTECT_WITH_INDEX(int_vect = NEW_INTEGER(0), int_vect_pindex); for (int i = 0; i n_reps; ++i) { Rprintf( ** Iteration %d:\n, i + 1); if (full_alloc) INTEGER_POINTER(int_vect)[i] = i; else { // This works now! -- SET_LENGTH(int_vect, GET_LENGTH(int_vect) + 1); REPROTECT(int_vect, int_vect_pindex); INTEGER_POINTER(int_vect)[GET_LENGTH(int_vect) - 1] = i; } } SEXP out, names, cls; PROTECT(out = NEW_LIST(1)); SET_VECTOR_ELT(out, 0, int_vect); PROTECT(names = NEW_CHARACTER(1)); SET_STRING_ELT(names, 0, COPY_TO_USER_STRING(integer.vector)); SET_NAMES(out, names); PROTECT(cls = NEW_CHARACTER(1)); SET_STRING_ELT(cls, 0, COPY_TO_USER_STRING(pushback)); classgets(out, cls); UNPROTECT(6); return out; } /CPP Code R Code nreps=5 allocate=FALSE sink(pushback_test.txt) test.pushback=.External(R_SimplePushBackTest, as.integer(nreps), as.logical(allocate)) print(test.pushback) sink() /R Code __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] Different scales for keys in lattice levelplots
On Friday 19 December 2003 14:23, Waichler, Scott R wrote: Can anyone tell me how to obtain a custom scale for the colorkey in each levelplot of a lattice? I am using lattice and levelplot to plot z = f(x, y) for multiple z variables. x and y values are the same across plots, but units for z are different and therefore I need to create a custom scale for each plot's key. It follows that I need to place each plot's key inside that plot's box. Should I use filled.contour() and gridBase instead? I'm not sure I understand your question. Are you using different calls to levelplot() for the different z variables ? What's wrong with the default colorkey for each such plot ? In any case, the arguments controlling colorkey are described in the help page for levelplot. Another way to control the scales (to be same across different calls to levelplot, say) is via the 'at' argument to levelplot (this can be used in a manner similar to what might have been expected from a 'zlim' argument, had one been available). Hope that helps, Deepayan __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] weighted regression
Paul Thanks for your quick response. Yes I have x and y backward. Thomas Blackwell pointed that out to me. Oops. And yes I will be considering more than 3 points for any regression. This little data set was first presented by a math wizard in a MathCAD forum explaining the mathematics behind weighted regression. When I tried Statistica on the data set I found that their definition of weights was a replication counter NOT in the spirit of 1/var. It put me into a blue funk ... I hate it when my state-of-the-art tools turn out to be a sham. I then turned to R and all appears to be mathemagically working well. As I walk the dogs I'm thinking (amittedly confusedly) on how prediction and confidence intervals could be generated around a weighted regression line. My thoughts range from purloining concepts from kriging, or bootstapping ... in each case I come to the conclusion that this is more complex topic than I thought Oh well...it keeps me interested in talking to smart people like you. Have a Merry Christmas REX - Original Message - From: Paul E. Johnson [EMAIL PROTECTED] To: [EMAIL PROTECTED] [EMAIL PROTECTED] Sent: Friday, December 19, 2003 2:43 PM Subject: Re: [R] weighted regression I keep new user tips at http://www.ku.edu/~pauljohn/R/Rtips.html I have some observations on your example: 1. You have y and x backward in your lm code 2. You get rid of the intercept by lm(y~x-1) but it is not generally advisable. it will bias the slope parameter if your assumption is wrong. 3. You are right that Weighted Least Squares can be accomplished by weights in lm. Please read the description of weight in ?lm weights: an optional vector of weights to be used in the fitting process. If specified, weighted least squares is used with weights 'weights' (that is, minimizing 'sum(w*e^2)'); otherwise ordinary least squares is used. If you understand WLS, then you will see how to enter the variable w. 4. I hope you aren't really doing regression with 3 observations. pj [EMAIL PROTECTED] wrote: To all I have some simple questions pertaining to weights used in regression. If the variability of the dependent variable (y) is a function of the magnitude of predictor variable (x), can the use of weights give an appropriate answer to the regression parameters and the std errors? Assume that y at x=1 and 6 has a standard deviation of 0.1 and at x=11 it is 0.4 Then according to a web page on weighted regression for a calibration curve at http://member.nifty.ne.jp/mniwa/rev006.htm, I should use 1/(std^2) for each weight. i.e. for x=1 and 6, w = 100 and x=11, w = 6.25 In R the run is: y-c(1,6,11) x-c(6.7,6.7,6.6) w-c(100,100,6.25) reg -lm(x~y, weight=w) summary(reg) Call: lm(formula = x ~ y, weights = w) Residuals: 123 -0.04762 0.09524 -0.19048 Coefficients: Estimate Std. Error t value Pr(|t|) (Intercept) 6.707619 0.025431 263.762 0.00241 ** y -0.002857 0.005471 -0.522 0.69361 --- Signif. codes: 0 `***' 0.001 `**' 0.01 `*' 0.05 `.' 0.1 ` ' 1 Residual standard error: 0.2182 on 1 degrees of freedom Multiple R-Squared: 0.2143, Adjusted R-squared: -0.5714 F-statistic: 0.2727 on 1 and 1 DF, p-value: 0.6936 Am I using the weight method correctly? And if so does the Estimated Std. Error for the Intercept and slope make sense? On another note. How does one do a regression with the origin fixed at 0? Merry Christmas REX [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help -- Paul E. Johnson email: [EMAIL PROTECTED] Dept. of Political Sciencehttp://lark.cc.ukans.edu/~pauljohn University of Kansas Office: (785) 864-9086 Lawrence, Kansas 66045FAX: (785) 864-5700 __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] error bars around a point
Andy Thank you. Downloaded Hmisc using the Packages button in the RGUI for Windows. My hat's off to the programmers... the system of downloading and installing a package works like a champ. I then tried your code and it too works great. I've added your suggested http://cran.r-project.org/search.html to my Favorites list. Very helpful. Merry Christmas REX Message: 36 Date: Fri, 19 Dec 2003 17:31:52 -0700 From: Andy Bunn [EMAIL PROTECTED] Subject: RE: [R] error bars around a point To: '[EMAIL PROTECTED]' [EMAIL PROTECTED], [EMAIL PROTECTED] Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=us-ascii See errbar in Hmisc. I found this with: help.search(error bar) And it's in the html help too. And on the r site: http://cran.r-project.org/search.html http://www.google.com/search?q=error+bardomains=r-project.orgsitesearc h=r-project.org # Example require(Hmisc) set.seed(1) x - 1:10 y - x + rnorm(10) delta - runif(10) errbar( x, y, y + delta, y - delta ) HTH, Andy [[alternative HTML version deleted]] __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Huynh-Feldt correction
Dear All, I tried to help myself by searching in newsgroups etc. I found a test for epsilon in a newsgroup (see attachement). I wrote a little skript to update the summary.anova if epsilon 0.75. Although I doublechecked my statistic book for the formula, the result is always the half of the epsilon calculated by SPSS. It is really strange. If I multiply espilon by two, I get the right results... regards, WP #begin - correctSummary - function(sum,eps) { # eps is the epsilon-value returned by function hf # sum is the summary.aov object in which the Huynh-Fldt correction has to be down len - length(sum) newSum - sum for (i in 2: len) { newSum[[i]][[1]][4,] - newSum[[i]][[1]][2,] newSum[[i]][[1]][3,] - newSum[[i]][[1]][2,] newSum[[i]][[1]][2,] - newSum[[i]][[1]][1,] rownames(newSum[[i]][[1]]) - c(rownames(newSum[[i]][[1]][1,]), Huynh-Feldt, rownames(newSum[[i]][[1]][2,]), Huynh-Feldt) newSum[[i]][[1]][2,][1] - newSum[[i]][[1]][2,][1]*epsilon*2 newSum[[i]][[1]][4,][1] - newSum[[i]][[1]][4,][1]*epsilon*2 newSum[[i]][[1]][2,][3] - newSum[[i]][[1]][2,][2]/newSum[[i]][[1]][2,][1] newSum[[i]][[1]][4,][3] - newSum[[i]][[1]][4,][2]/newSum[[i]][[1]][4,][1] newSum[[i]][[1]][2,][4] - newSum[[i]][[1]][2,][3]/newSum[[i]][[1]][4,][3] newSum[[i]][[1]][2,][5] - pf(as.numeric(newSum[[i]][[1]][2,][4]), as.numeric(newSum[[i]][[1]][2,][1]), as.numeric(newSum[[i]][[1]][4,][1]), lower.tail = FALSE) } newSum } hf - function(m){ # m is a matrix with subjects as rows and conditions as columns # note that checking for worst case scenarios F correction first might # be a good idea using J/(J-1) as the df correction factor n- length(m[,1]) J-length(m[1,]) X-cov(m) r- length(X[,1]) D-0 for (i in 1: r) D- D+ X[i,i] D-D/r SPm- mean(X) SPm2- sum(X^2) SSrm-0 for (i in 1: r) SSrm- SSrm + mean(X[i,])^2 epsilon - (1/(J-1))*((J^2*(D-SPm)^2)/(SPm2 - 2*J*SSrm + J^2*SPm^2)) epsilon } # On Thursday December 18 2003 12:00, Wolfgang Pauli wrote: Dear R-helpers, Does anybody know, whether there is an option to tell aov/anova, or do something similar, to get a Huynh-Feldt correction of dfs in the aov/anova function? Thanks in advance! __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Software/algorithms for competing risks analysis
Dear R group: I am compiling a list of available algorithms/macros/software for performing the following types of competing risks analyses: (1) cumulative (crude) incidence regression (2) Gray's K-sample log-rank test for cmulative incidence (3) multi-state models (4) dependent competing risks modeling with (a) copulas, (b) Robin's Inverse probability of censoring weighted (IPCW) estimator, (c) frailty I would really appreciate any pointers/references to algorithms/software in other environments such as SAS/STATA/SPSS, as well as in R. Thanks very much, Ravi. __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] Sweave/LaTeX Problem with EPS PDF
Dear List: I am unsure if my problem is with Sweave or LaTeX. Anyhow, I am using the MikTeX distribution and TexnicCenter. I can easily create Sweave files and all goes well until I try to incorporate graphics. I use the same code as found in the examples found in the users manual. In R, the graphics I want are created as Sweave is creating the .tex file. When I examine the .tex file created by Sweave, it includes the includegraphics{} statement needed. When I run LaTeX on the .tex file, everything works except that the graphics I want are not displayed. However, when I examine the pdf or eps files created, there is nothing there. When I view the EPS using Ghostview, the file is empty, but there appears to be a bounding box surrounding nothing. When I open the pdf graphic, there is nothing there either. I have tried creating both dvis and pdf files. Again, text works perfectly, but graphics do not work. Does anyone have any suggestions? Many thanks, Harold __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
[R] varFixed
Dear List: Earlier this week I posted a question and received no response, and I continue to struggle with my model. My original question is pasted below. I am using lme and want to fix the variance of the within group residual at 1 (e~n(0,1). I think the varFixed function should be used to accomplish this, but I am struggling to figure out how to do this. Can anyone offer suggestions on how this might be accomplished? Thanks, I would appreciate any suggestions. Harold Dear List: I am trying to figure out how to incorporate measurement error in an longitudinal educational data set using lme to create a true score model. As a by-product of the procedures used to scale educational tests, one can obtain a person-specific measurement error associated with each score, or a conditional standard error. For example, a score of 200 would have measurement error specific to that score that would be different than, say, a score of 250. I have been rather successful in figuring out how to rescale the necessary components to create this true score model. This simply requires that the response variable, the intercept, and any other variables in the design matrix be multiplied by the reciprocal of the standard error of measurement for the associated score. There may be a better way to do this, but I manually create a vector of 1s for all observations and multiply this vector by 1/sem. This is the new intercept. I also multiply any other predictors in the design matrix by the same value. In the R code, I remove the intercept included by default (-1) and include the newly created intercept (which is no longer a constant) as well as the new response variable and rescaled predictors. However, I am confused regarding the within-group error term. Fitting this model requires that the variance be fixed at 1: e ~ n(0,1). Is it possible to constrain the variance for this model as such? I would appreciate any comments or suggestions regarding this model. __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
Re: [R] Sweave/LaTeX Problem with EPS PDF
Harold Doran wrote: [... Sweave use...] However, when I examine the pdf or eps files created, there is nothing there. When I view the EPS using Ghostview, the file is empty, but there appears to be a bounding box surrounding nothing. When I open the pdf graphic, there is nothing there either. I have tried creating both dvis and pdf files. Again, text works perfectly, but graphics do not work. Are they lattice graphics? They need to be wrapped in a print() statement, like print(xyplot(...)) Without that, they don't produce output to eps or pdf. Cheers Jason -- Indigo Industrial Controls Ltd. http://www.indigoindustrial.co.nz 64-21-343-545 [EMAIL PROTECTED] __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help