Re: [R] Generalized Estimating Equations and log-likelihood calculation

At 18:29 2004-02-18, you wrote: Is anyone aware of a way to calculate log-likelihood for GEE models? No. GEE fitting is based on quasi-likelihood. However, it is possible to derive AIC-like measures based on the quasi-likelihood. Lebreton et al (1992) suggested a simple adjustment in the GLM

Re: [R] Generalized Estimating Equations and log-likelihood calculation

At 09:17 2004-02-19, you wrote: For GEE models, Pan (2001) has introduced QIC. None of these measures are implemented in R or in any add-on package as far as I know. Actually, take a look at http://hisdu.sph.uq.edu.au/lsu/SSAI%20course/course_tools.htm Henric

[R] Setting ylim while plotting an 'its' object

Hi, I have an object 'sat' of class 'its'. How do I have to set the ylim parameter to change the y Axis limits while plotting a time series? Thanks very much for your help, Marcel class(sat) [1] its attr(,package) [1] its range(sat) [1] -0.5908360386 0.0001541759

Re: [R] Summary - normality test

Thanks to everyone for the responses! Samuel SUMMARY : MY QUESTION : Hello, I am analysing several samples whose sizes are from 9 to 110. I would like to test their distribution with R, whether they are normal or not. I wonder which test for normality from R should I use . Thank you for

RE: [R] pass by reference -- how to do it

For the record, be careful: - does not necessarily assign to the global environment. In R ` x - value` assigns `value` to the first instance of `x` it can find using lexical scoping. Only if it doesn't find any such variable, it will indeed create an `x` in .GlobalEnv. Tricky for those

Re: [R] Setting ylim while plotting an 'its' object

It looks like you need to set yrange, not ylim: see ?itsDisp. I don't see a good reason for this. On Thu, 19 Feb 2004, Marcel Sutter wrote: Hi, I have an object 'sat' of class 'its'. How do I have to set the ylim parameter to change the y Axis limits while plotting a time series?

[R] How to create a nb object?

Hi group, I'd like to do spatial analysis of my data using the spdep package. It appears that a file of class nb is necessary, but I do not find what that should look like and if there is a function that creates such file for me. How can I create a nb-object of my data points? Thanks, René

Re: [R] How to create a nb object?

On Thu, 19 Feb 2004, René Eschen wrote: Hi group, I'd like to do spatial analysis of my data using the spdep package. It appears that a file of class nb is necessary, but I do not find what that should look like and if there is a function that creates such file for me. If you have points

[R] suppressing non-integer labels for plot x-axis

Dear R-helpers, I am having difficulty making R plot only integer labels on the x-axis of a simple graph. I want to plot the median values of a score on each of three occasions. Non-integer occasions are impossible. But, R keeps labelling the x-axis with half-occasions, despite my attempts to stop

[R] F Dist

Hi, why is it that in tables for the F distribution it is for example F0.95[6,28]=2.45 or F0.95[10,20]=2.35 but pf(6,28,2.45) [1] 0.8854934 pf(10,20,2.35) [1] 0.9300167 Thanks for clarifying, jeorg __ [EMAIL PROTECTED] mailing list

RE: [R] F Dist

... because that should have been: pf(2.45, 6, 28) [1] 0.9503618 pf(2.35, 10, 20) [1] 0.950176 When arguments are not named, the order is significant. Andy From: Joerg Schaber Hi, why is it that in tables for the F distribution it is for example F0.95[6,28]=2.45 or

Re: [R] F Dist

Hello, the order of argument is not the correct one: p(2.45,6,28) [1] 0.9503618 P.A. On Thu, 19 Feb 2004, Joerg Schaber wrote: Hi, why is it that in tables for the F distribution it is for example F0.95[6,28]=2.45 or F0.95[10,20]=2.35 but pf(6,28,2.45) [1] 0.8854934

Re: [R] suppressing non-integer labels for plot x-axis

Jonathan Williams wrote: Dear R-helpers, I am having difficulty making R plot only integer labels on the x-axis of a simple graph. I want to plot the median values of a score on each of three occasions. Non-integer occasions are impossible. But, R keeps labelling the x-axis with half-occasions,

R for economists (was: [R] Almost Ideal Demand System)

Hi, I did not find any web page about using R in economics and econometrics so far. However, this does not mean that there is none (searching with google for R and economics gives many pages about economics and a name like Firstname R. Lastname on it ;-)). Does anybody in the list does know

RE: [R] pass by reference -- how to do it

[sorry if this appears twice. I had an email problem and am sending it out again] As you and Andy correctly point out, - searches through its environments and it may find a match prior to the Global Environment. On the other hand, while its possible to get into trouble, I believe its actually

Re: R for economists (was: [R] Almost Ideal Demand System)

Hello, I know a few papers in economics and econometrics using R. One of them in the Journal of Applied Econometrics. You may have a look at the following link: http://netec.mcc.ac.uk/WoPEc/data/Articles/jaejapmetv:14:y:1999:i:3:p:319-29 .html or you could download the paper from the R-project

[R] polr warning message optim

Hello R-users, I am using polr function in library(MASS). The code I use is the following: polr(as.ordered(q23p)~.,data=as.data.frame(datapr2)) where datapr2 is a matrix of 63 columns (together with the dependent variable) and 1665 rows. But I am receiving the warning message Error in

RE: [R] pass by reference -- how to do it

I don't disagree - it's just that I once spent a day trying to work out why my old S-Plus library's - didn't work in R as I thought it should - and then I read the FAQ ... Especially in interactive work, - is very unlikely to cause a problem. But it's definitely dodgy to trust to this idiom

[R] reshape direction=wide

Hello I am reshaping a data.frame bids -- reshaped as shown below. I thought this should be possible with a single invocation of reshape, but the only way I came up with is reshaping subsets for each keyword and then joining them together. Does anyone have an idea how to solve this in a more

[R] Obtaining SE from the hessian matrix

Dear R experts, In R-intro, under the 'Nonlinear least squares and maximum likelihood models' there are ttwo examples considered how to use 'nlm' function. In 'Least squares' the Standard Errors obtained as follows: After the fitting, out$minimum is the SSE, and out$estimates are the

Re: R for economists (was: [R] Almost Ideal Demand System)

On Thu, Feb 19, 2004 at 02:26:55PM +0100, Arne Henningsen wrote: Hi, I did not find any web page about using R in economics and econometrics so far. However, this does not mean that there is none (searching with google for R and economics gives many pages about economics and a name like

Re: [R] suppressing non-integer labels for plot x-axis

On Thu, 2004-02-19 at 07:20, Jonathan Williams wrote: Dear R-helpers, I am having difficulty making R plot only integer labels on the x-axis of a simple graph. I want to plot the median values of a score on each of three occasions. Non-integer occasions are impossible. But, R keeps labelling

RE: [R] ANOVA procedure on the sufficient statistics

You can apply the trick that Prof. Dalgaard recently posted in response to a similar question (for one-way ANOVA). For each cell, generate data as: y - cell.mean + cell.sd * scale(rnorm(cell.count)) Then generate the data frame to feed to aov. That will generate only approximately the same

[R] More variables on pca

Hi, When I have more variables than units, e.g. a 10x20 data frame with 20 variables. princomp can't do pca on variables. prcomp from mva package can do th etrick but gave 10 principal components. Should we expect 20 for this case? Thanks in advance, Weidong

RE: R for economists (was: [R] Almost Ideal Demand System)

Hello, -Original Message- From: Mahmoud K. Okasha [SMTP:[EMAIL PROTECTED] Sent: Thursday, February 19, 2004 3:14 PM To: [EMAIL PROTECTED] Subject: Re: R for economists (was: [R] Almost Ideal Demand System) Hello, I know a few papers in economics and econometrics using R.

[R] filling the area between two curves in a plot

Hi, does anybody know how I can color the area enclosed between two curves on a plot? Thanks, --- Rajarshi Guha [EMAIL PROTECTED] http://jijo.cjb.net GPG Fingerprint: 0CCA 8EE2 2EEB 25E2 AB04 06F7 1BB9 E634 9B87 56EE

Re: R for economists (was: [R] Almost Ideal Demand System)

On Thu, 19 Feb 2004 16:14:17 +0200 Mahmoud K. Okasha wrote: Hello, I know a few papers in economics and econometrics using R. One of them in the Journal of Applied Econometrics. You may have a look at the following link:

[R] Re: R for economists

Arne Henningsen wrote: Hi, I did not find any web page about using R in economics and econometrics so far. However, this does not mean that there is none (searching with google for R and economics gives many pages about economics and a name like Firstname R. Lastname on it ;-)). Does anybody

Re: R for economists (was: [R] Almost Ideal Demand System)

On Thu, 19 Feb 2004 19:06:08 +0530, Ajay Shah [EMAIL PROTECTED] wrote: snip R for economists Okay, I made a start, with http://www.mayin.org/ajayshah/KB/R/more.html Tell me what should go into it. :-) I think the most important thing for economists would be a list of the common

RE: [R] Area between CDFs

On Wed, 18 Feb 2004, Samuelson, Frank* wrote: You may not want to integrate cdfs. They're already probabilities. :) Nice analytic statistics exist for just the maximum distance between the cdfs, for example. And for the area between cdfs, which is perhaps better known as the difference in

RE: [R] filling the area between two curves in a plot

Assuming the coordinates for the two curves are x1, y1, x2 and y2, and that the pairs are sorted by x1 and x2, respectively. Then just do something like: polygon(c(x1, rev(x2)), c(y1, rev(y2), ...) HTH, Andy From: Rajarshi Guha Hi, does anybody know how I can color the area enclosed

[R] read.socket - Strange strings. How to force sub to remove all occurences of a pattern?

I am opening a connection to an apache server sending a get and then I am reading from the socket. All works finde except that I am getting some strange strings disrupting the html e.g. FON\r\nffb\r\nT COLOR ... \r\na48\r\n \r\nffb\r\n They are not frequent ( a few of them) but of course

Re: [R] NAs introduced by coercion warning?

Jonathan, It's still hard to tell. Try this: options(warn = 1) # see ?options for explanation ## RUN YOUR CODE Regards, Sundar Jonathan Greenberg wrote: Its hard for me to pinpoint where this is happening, since I'm working on an image that¹s about 1 x 2 pixels, and 12 bands deep and

Re: [R] ANOVA procedure on the sufficient statistics

On Thu, 19 Feb 2004 08:45:01 -0500, Michael Friendly [EMAIL PROTECTED] wrote : You can apply the trick that Prof. Dalgaard recently posted in response to a similar question (for one-way ANOVA). For each cell, generate data as: y - cell.mean + cell.sd * scale(rnorm(cell.count)) Then generate

[R] Comparing two regression slopes

I would suggest the method of Sokal and Rholf (1995) S. 498, using the F test. Below I repeat the analysis by Spencer Graves: Spencer: df1 - data.frame(x=1:3, y=1:3+rnorm(3)) df2 - data.frame(x=1:3, y=1:3+rnorm(3)) fit1 - lm(y~x, df1) s1 - summary(fit1)$coefficients fit2 - lm(y~x, df2) s2

Re: [R] filling the area between two curves in a plot

Rajarshi Guha wrote: Hi, does anybody know how I can color the area enclosed between two curves on a plot? This is possible with the polygon function. Try something like: # create some data x - sort(runif(10, min=0, max=10)) y - runif(10, min=2, max=5) #Polygon-Plot plot(x,y, type=n,

Re: [R] building the development version

Deborah Swayne [EMAIL PROTECTED] writes: Peter Dalgaard writes: OK, that I can confirm. They do turn up on a checkout without the -P option, but disappear after cvs up -Pd. Tony, Debbie, did you perhaps configure before attempting to prune the empty dirs? Could you try

Re: [R] filling the area between two curves in a plot

On Thu, 19 Feb 2004, Rajarshi Guha wrote: Hi, does anybody know how I can color the area enclosed between two curves on a plot? There is an example of this in demo(graphics). -thomas __ [EMAIL PROTECTED] mailing list

RE:[R] reshape direction=wide

v.names=c(var1,var2) creates a separate column for each combination of variables in v.names and level of variable identified by timevar. I am reshaping a data.frame bids -- reshaped as shown below. I thought this should be possible with a single invocation of reshape, but the only way I came

Re: [R] Obtaining SE from the hessian matrix

On Thu, 19 Feb 2004, Timur Elzhov wrote: So, what is the _right_ way for obtatining SE? Why two those formulas above differ? If you are maximising a likelihood then the covariance matrix of the estimates is (asymptotically) the inverse of the negative of the Hessian. The standard errors are

RE: [R] read.socket - Strange strings. How to force sub to remove all occurences of a pattern?

Hi, here's some classical questions to you; what's your system and what version of R are you using? Could you give a workable code example how you connect to the Apache server? Is the server public so others can test your code? About sub(): you're looking for gsub(). Cheers Henrik Bengtsson

[R] efficient matrix approx

Hello, I am looking for a highly efficient matrix version of linear interpolation (like approx). As an example I have data like follows: x-data.frame(time=1:20, x=(1:20)/10, y=runif(20)) t - seq(1.5, 15.5 ,by=0.5) # and I found the following solution: nam - names(x) app -

Re: [R] Comparing two regression slopes

I missed the earlier thread, but if I had the data and not just the table of coefficients and standard errors, then I'd try combining the data sets as follows: set.seed(1) df1 - data.frame(x=1:3, y=1:3+rnorm(3)) df2 - data.frame(x=1:3, y=1:3+rnorm(3)) DF. -

Re: [R] read.socket - Strange strings. How to force sub to remove all occurences of a pattern?

Hallo Douglas, Henrik! Thanks a lot for the help with sub and gsub. Eryk. *** REPLY SEPARATOR *** On 2/19/2004 at 10:12 AM Douglas Grove wrote: Did you look at the help page for sub?? Read Details: On Thu, 19 Feb 2004, wolski wrote: I am opening a connection to an

RE: R for economists (was: [R] Almost Ideal Demand System)

Not exactly R-related, but since there seems to be a lot of interest in this topic, I'd also mention the book: Modeling Financial Time Series with S-PLUS, by Eric Zivot et al, which has several chapers on econometric modeling with S. The book website is at:

Re: [R] building the development version

Peter Dalgaard [EMAIL PROTECTED] writes: Deborah Swayne [EMAIL PROTECTED] writes: Peter Dalgaard writes: OK, that I can confirm. They do turn up on a checkout without the -P option, but disappear after cvs up -Pd. Tony, Debbie, did you perhaps configure before attempting to

RE: [R] Area between CDFs

Thanks. I have been able to create the following simple function to examine the vertical gap between two CDFs at a value along the x-axis that I specify. For example, I create the ECDFs: male.ecdf-ecdf(egmale$math) female.ecdf-ecdf(egfemale$math) I then define the following function:

RE:[R] reshape direction=wide

On Thu, 19 Feb 2004, Hanke, Alex wrote: v.names=c(var1,var2) creates a separate column for each combination of variables in v.names and level of variable identified by timevar. yes, but that's not what he actually wants. He needs to create a new id variable that combines ID and BID first

Re: [R] Obtaining SE from the hessian matrix

Minor correction: Most likely, Prof. Lumley's statement is correct. However, as I'm sure he knows, it depends on what you are maximizing or minimizing: If you are maximizing the log(likelihood), then the NEGATIVE of the hessian is the observed information. This latter should be

[R] R won't start

Hi, I have a problem with R; it won't start! I've installed it fine, no problems, but when I try to run it the mouse cursor changes to an hourglass for 1 second and back to normal, and that's it, it doesn't start up. No error messages or nothing. I run Windows XP Pro. by the way. I'd be

[R] solving equations with several variables

I am relatively new to R, but I would like to use it find the values of b1 and b2 that maximize the following equation: (sum(cos(x3+2.33474-2(atan(b1*x1+b2*x2) Any help would be greatly appreciated. __ [EMAIL PROTECTED] mailing list

RE: [R] R won't start

Hi, -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Pål Anders Martinussen Sent: Friday, February 20, 2004 8:53 AM To: [EMAIL PROTECTED] Subject: [R] R won't start Hi, I have a problem with R; it won't start! I've installed it fine, no

Re: [R] efficient matrix approx

Thomas Petzoldt [EMAIL PROTECTED] writes: Hello, I am looking for a highly efficient matrix version of linear interpolation (like approx). As an example I have data like follows: x-data.frame(time=1:20, x=(1:20)/10, y=runif(20)) t - seq(1.5, 15.5 ,by=0.5) # and I found the following

Re: [R] solving equations with several variables

optim? Josh Schmidt wrote: I am relatively new to R, but I would like to use it find the values of b1 and b2 that maximize the following equation: (sum(cos(x3+2.33474-2(atan(b1*x1+b2*x2) Any help would be greatly appreciated. __ [EMAIL

Re: [R] solving equations with several variables

Josh Schmidt wrote: I am relatively new to R, but I would like to use it find the values of b1 and b2 that maximize the following equation: (sum(cos(x3+2.33474-2(atan(b1*x1+b2*x2) Any help would be greatly appreciated. Given b1 and b2 are scalars (we are minimizing f, hence

Re: [R] R won't start

On Thu, 19 Feb 2004 20:52:50 +0100, Pål Anders Martinussen [EMAIL PROTECTED] wrote : Hi, I have a problem with R; it won't start! I've installed it fine, no problems, but when I try to run it the mouse cursor changes to an hourglass for 1 second and back to normal, and that's it, it doesn't

RE: [R] solving equations with several variables

I tried: f-function(b,x)-sum(cos(x1+2.33474-2*(atan(b[1]*x2+b[2]*x3 optim(c(0, 0), f, x1=x1, x2=x2, x3=x3) and got this error. Error in fn(par, ...) : unused argument(s) (x1 ...) I am trying to solve for the circular-linear regression equation in Fisher (1993). = Original Message

[R] controlling nls errors

Hello. I am using the nonlinear least squares function (nls). The function that I am trying to fit seems to be very sensitive to the starting values and, if these are not chosen properly, the nls function stops and gives an error message: Error in numericDeriv(form[[3]], names(ind), env) :

[R] lme problem

Exactly what rule has been broken when lme (package nlme) produces the error message Incompatible formulas for groups in random and correlation? Here is what I am trying to do: lme(Y ~ trait-1, random = ~trait-1|sire, corr = corSymm(form = ~trait|cow)) Trait, sire and cow are factors (cow

RE: [R] importing ascii grids (for gstat)

No problem. I'm glad it worked. It would be pretty easy to package as a function for gstat. -Andy -Original Message- From: femke [mailto:[EMAIL PROTECTED] Sent: Thursday, February 19, 2004 2:54 PM To: Andy Bunn Subject: Re: [R] importing ascii grids (for gstat) Thanks very

[R] 1024GB max memory on R for Windows XP?

I have 2GB installed on my windows XP box running R 1.9.0, and after performing a prune.tree(intree,newdata), I get an out of memory error within R, but it says the maximum allowed is 1024gb (1/2 of what I have!) Can R not use more than 1GB on an XP box? I noticed I had ~600mb left over after R

[R] piece wise application of functions

Dear all, After struggling for some time with *apply() and eva() without success, I decided to ask for help. I have 3 lists labeled with, each contains 3 different interpolation functions with identical names: names(missgp0) [1] spl.1mb spl.2mb spl.5mb names(missgp1) [1] spl.1mb spl.2mb

Re: [R] 1024GB max memory on R for Windows XP?

See ?Memory. Setting memory.limit(2048) in R may help -roger Jonathan Greenberg wrote: I have 2GB installed on my windows XP box running R 1.9.0, and after performing a prune.tree(intree,newdata), I get an out of memory error within R, but it says the maximum allowed is 1024gb (1/2 of what I

[R] Does pdf() not work on Trellis graphics?

Hello, I have the following example. testpdf - function(x,y){ pdf() #plot(x,y) xyplot(x~y) dev.off() } source(testpdf.R) x=1:1000 testpdf(x,x) with xyplot, the output is a file that can not be read because it has no pages. With plot it works fine. If I do it manually (from the GUI

Re: [R] 1024GB max memory on R for Windows XP?

On Thu, 19 Feb 2004 15:14:36 -0800, Jonathan Greenberg wrote: I have 2GB installed on my windows XP box running R 1.9.0, and after performing a prune.tree(intree,newdata), I get an out of memory error within R, but it says the maximum allowed is 1024gb (1/2 of what I have!) Can R not use more

[R] Using R remotely on a Mac OS X machine

Hello, I've recently installed R (RAqua I guess?) on a Mac OS X machine and it works fine locally. However, I want to run some things from the command line remotely. I can't seem to get this to work and have tried several different things (i.e., --vanilla, --gui none, doing stuff in batch

[R] Sweave not found from MikTeX?

I've been working on a LaTeX document that contains Sweave code and cannot get MikTeX to find the Sweave.sty file. I've added the c:\rw1081\share\texmf path in the MikTeX roots (I've ven added the path in the environment variables ) but to no avail. Is there a trick to getting Sweave installed

Re: [R] Does pdf() not work on Trellis graphics?

See the FAQ, 7.24. you need to print(xyplot(x~y)) Hello, I have the following example. testpdf - function(x,y){ pdf() #plot(x,y) xyplot(x~y) dev.off() } source(testpdf.R) x=1:1000 testpdf(x,x) with xyplot, the output is a file that can not be read because it has no

Re: [R] 1024GB max memory on R for Windows XP?

Hmm, I went over to a unix box that should have had plenty of RAM and swap space and running the same prune.tree process ended up causing R to be Killed (e.g. It quit out). Thoughts? --j On 2/19/04 4:58 PM, Duncan Murdoch [EMAIL PROTECTED] wrote: On Thu, 19 Feb 2004 15:14:36 -0800, Jonathan

Re: [R] 1024GB max memory on R for Windows XP?

Was this by any chance a Linux on x86? If so, I think there's roughly a 2.5-3GB limit for a single process imposed by the architecture. -roger Jonathan Greenberg wrote: Hmm, I went over to a unix box that should have had plenty of RAM and swap space and running the same prune.tree process

[R] [R-pkgs] New Package: multinomRob

We would like to announce the availability on CRAN of a new package multinomRob. It does robust estimation of overdispersed multinomial regression models. The package is also able to estimate overdispersed grouped multinomial logistic and multivariate-t logistic models. The code is relatively

[R] Confidence intervals for logistic regression

Hi, I found myself trying to figure out the type of confidence interval used for the coefficients of the logistic regression fit by using glm(family=binomial)... I suspect it is Wald confidence interval but am not sure...Does anybody know? Also, if so, how can I ask for likelihood ratio and/or

[R] (no subject)

__ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

Re: [R] solving equations with several variables

Josh Schmidt wrote: I tried: f-function(b,x)-sum(cos(x1+2.33474-2*(atan(b[1]*x2+b[2]*x3 optim(c(0, 0), f, x1=x1, x2=x2, x3=x3) Sorry, should have been: f - function(b, x1, x2, x3){ -sum(cos(x1+2.33474-2*(atan(b[1]*x2+b[2]*x3 } optim(c(0, 0), f, x1=x1, x2=x2, x3=x3) Uwe Ligges

Re: [R] Confidence intervals for logistic regression

On Thu, 19 Feb 2004, Michael Levine wrote: I found myself trying to figure out the type of confidence interval used for the coefficients of the logistic regression fit by using glm(family=binomial)... AFAIK, no confidence interval is given by that call: it does not even calculate standard