similar to: Compatibility of external C code across platforms

> On Apr 10, 2016, at 3:11 AM, Murray Efford <murray.efford at otago.ac.nz> wrote: > > Martin - > Thanks, but although hatvalues() is useful for calculating PRESS, I can't find anything directly relevant to my question in the influence help pages. After some burrowing in the literature I'm doubting there is an answer out there (PRESS R^2 is always presented in a fairly

Martin - Thanks, but although hatvalues() is useful for calculating PRESS, I can't find anything directly relevant to my question in the influence help pages. After some burrowing in the literature I'm doubting there is an answer out there (PRESS R^2 is always presented in a fairly ad hoc way). This is a new topic, as you say, and perhaps better handled on a statistics list. Murray Efford

> On Apr 10, 2016, at 9:38 AM, David Winsemius <dwinsemius at comcast.net> wrote: > >> >> On Apr 10, 2016, at 3:11 AM, Murray Efford <murray.efford at otago.ac.nz> wrote: >> >> Martin - >> Thanks, but although hatvalues() is useful for calculating PRESS, I can't find anything directly relevant to my question in the influence help pages. After

>>>>> Murray Efford <murray.efford at otago.ac.nz> >>>>> on Fri, 8 Apr 2016 18:45:33 +0000 writes: > Thanks for these perfectly consistent replies - I didn't > understand the purpose of m = sum(w * f/sum(w)) and saw it > merely as a weighted average of the fitted values. My > ultimate concern is how to compute an appropriate

Thanks for these perfectly consistent replies - I didn't understand the purpose of m = sum(w * f/sum(w)) and saw it merely as a weighted average of the fitted values. My ultimate concern is how to compute an appropriate weighted TSS (or equivalently, MSS) for PRESS-R^2 = 1 - PRESS/TSS = 1 - PRESS/ (MSS + PRESS). Do you think it then makes sense to substitute the vector of leave-one-out fitted

Dear R users I would like to simulate underdispersed Poisson and binomial distributions somehow. I know you can do this for overdispersed counterparts - using rnbinom() for Poisson and rbetabinom() for binomial. Could anyone share functions to do this? Or please share some tips for modifying existing functions to achieve this. Thank you very much for your help and time Shinichi

On 08 Apr 2016, at 12:57 , Duncan Murdoch <murdoch.duncan at gmail.com> wrote: > On 07/04/2016 5:21 PM, Murray Efford wrote: >> Following some old advice on this list, I have been reading the code for summary.lm to understand the computation of R-squared from a weighted regression. Usually weights in lm are applied to squared residuals, but I see that the weighted mean of the

I have just used the RAqua.pkg to install R, and it doesn't work! I have Mac OS X 10.3, and the old version of R works OK, but the .pkg thing installed, and then when I double-click on the "StartR" application nothing happens. Nothing at all, no little icon bobbing up and down in the dock, just nothing. What have I done wrong? I didn't have to install the other packages

Hi, I am trying to connect points on a graph that originate from *different columns of data*. For each sample I have minimum and maximum data points and I would like to draw a line connecting these in order to visualize the spread, as well as where each sample is in relation to the x-axis. So far I can generate the points, but the only lines I have been able to make join all the minimum values

Dear All I'm logging the stdout and stderr of an R program into two separate files (stderr.txt and stdout.txt) using sink() I would like to append extra information such as "date", "memory usage" etc to every line of output that goes to stdout or stderr. For example > cat("hello \n") should give output that looks something like: "hello --- Thu Jun 19

Why is 19 March, 1947 a little longer than one day? x <- as.POSIXct("1947-04-16") julian(x, origin = as.POSIXct("1947-03-20")) Time difference of 27 days julian(x, origin = as.POSIXct("1947-03-19")) Time difference of 28.04167 days > julian(x, origin = as.POSIXct("1947-03-18")) Time difference of 29.04167 days I am running R-1.7.1 compiled on

Do you mean w <- z$residuals ? Type names(z) to see the list of item in your model. I ran your code on a lm and it work fine. You don't need the brackets around mss <- Michael Long On 04/07/2016 02:21 PM, Murray Efford wrote: > Following some old advice on this list, I have been reading the code for summary.lm to understand the computation of R-squared from a weighted

Following some old advice on this list, I have been reading the code for summary.lm to understand the computation of R-squared from a weighted regression. Usually weights in lm are applied to squared residuals, but I see that the weighted mean of the observations is calculated as if the weights are on the original scale: [...] f <- z$fitted.values w <- z$weights [...] m

I construct a pdf package manual in Windows 7 using R CMD Rd2dvi --pdf --no-preview [packagename] Page numbers are listed correctly under 'R topics documented' at the front, but incorrectly (offset by -2 pages) in the Index at the back. Following the hyperlinked page numbers in the Index takes you to the wrong page. 2 pages happens to be the length of the package overview man page inserted

Dear All I recently proposed a simple modification to Wedderburn's 1974 estimate of overdispersion for count and binomial data, which is used in glm for the quasipoisson and quasibinomial families (see the reference below). Although my motivation for the modification arose from considering sparse data, it will be almost identical to Wedderburn's estimate when the data are not sparse.

Hi there folks I have R 1.3.0 running on Red Hat Linux 7.1 with GhostView 3.5.8. When I produce a file like: > bitmap("foo.pdf", type="pdfwrite") > plot(foo) > dev.off() and then try to read it with gv, I get an unrecoverable error and nothing displayed. Any help as to whether this ***should*** work would be appreciated. cheers, John -- John Williams

On 07/04/2016 5:21 PM, Murray Efford wrote: > Following some old advice on this list, I have been reading the code for summary.lm to understand the computation of R-squared from a weighted regression. Usually weights in lm are applied to squared residuals, but I see that the weighted mean of the observations is calculated as if the weights are on the original scale: > > [...] > f

The following reproducable code shows the setting of my problem: set.seed(260808) n = 50 x = rnorm(n) y = rnorm(n) z = ceiling(runif(n,0,4)) g = runif(n,0,6) G = factor(ceiling(g)) xyplot(y ~ x | G) plsy <- trellis.par.get("plot.symbol") plsy$pch = z trellis.par.set("plot.symbol",plsy) xyplot(y ~ x | G) plsy$pch = as.character(z)

Hi, Is there any equivalent of type="n" when constructing barplots which will still construct the axes (plot=F, as it says doesn' plot anything at all). Alternatively I tried setting col="white" and border="white" but the border command does not seem to be operational. True?? Any other ideas? What I'm actually trying to do is construct vertical abline()'s

Hello, I am trying to run a PCA in R and I cannot get the PC scores for each of the values. Using pcX <- princomp(X) then loadings(pcX) I can get a listing of the eigenvectors but not the actual PC scores for each value in the dataset. I greatly appreciate any help anyone can offer Thanks Ken