similar to: About Omega in pda()

** High Priority ** I am a french student working on discriminant analysis issues. I am a new user of R language and have a few questions about its use : + About the mda package how can I plot the decision boundaries in the space of predictors from the results of the fda function ? + Does R include 3-D plots ? + How can I use bootstrap and cross-validation in classification ? I

Hi, how can i write this simple sentence : "Hello world" with "Hello" only in bold ? I try > plot(1:5) > title(main=paste(expression(bold("Hello")),"world",sep=" ")) but the result is wrong. thanks, Bruno Si vous n'etes pas destinataires de ce message, merci d'avertir l'expediteur de l'erreur de distribution et de le

Hello, with a data.frame like this : > toto <- data.frame(id=c("id1","id1","id2","id3","id3","id3"),dpt=c("13","13","34","30","30","30")) > toto id dpt 1 id1 13 2 id1 13 3 id2 34 4 id3 30 5 id3 30 6 id3 30 what is the most efficient ways to obtain : id

Dear R-users, i have two questions : 1- first of all, i wish to know the way to obtain a serie with a format like "00" : ( "01","02","03","04"....) or like postal code ("01100","02222"). for instance, i do : > format(strptime(as.character(c(1:4)),"%H"),"%H") but it sounds complicate and not really

Hi, I have been trying to fit an un-penalised gam in mgcv (in order to get more reliable p-values for hypothesis testing), but I am struggling to get the model to fit sucessfully when I add in a te() interaction. The model I am trying to fit is: gam(count~ s(x1, bs = "ts", k = 4, fx = TRUE) + s(x2, bs = "ts", k = 4, fx = TRUE) + te(x2, x3, bs =

Hi, i have two data.frame x and y like : > x <- data.frame( num = c(1:10), value = runif(10) ) > y <- data.frame( num = c(6:10), value = runif(5) ) and i want to obtain something like : num.x value.x num.y value.y 1 0.38423828 NA 0.2911089 2 0.17402507 NA 0.8455208 3 0.54443465 NA 0.8782199 4 0.04540406 NA 0.3202252 5 0.46052426

Good Afternoon Everyone, I am looking for advice fitting a linear mixed model where the random components do not seem to fit within the model formulae for lmer. The columns of Z are not stratified and have the notional random formula (z1 | 1) + ... + (zk | 1). Context I am fitting a penalised thin plate spline with knots k1 to kn. The basis functions Zk are |x-ki|^3 and the penalty matrix has

Anyone using the penalised partial likelihood routines in survival5 should update their version. A bug has been fixed in the S package: in coxph() models with penalised likelihood and strata it was possible in some circumstances to get an infinite loop or perhaps an incorrect answer. The new version (2.3) is on cran.r-project.org and will percolate through CRAN in the next few days. -thomas

Anyone using the penalised partial likelihood routines in survival5 should update their version. A bug has been fixed in the S package: in coxph() models with penalised likelihood and strata it was possible in some circumstances to get an infinite loop or perhaps an incorrect answer. The new version (2.3) is on cran.r-project.org and will percolate through CRAN in the next few days. -thomas

A nearly complete port of the new survival5 package has been sent to CRAN and will soon be appearing on a mirror near you in the contrib/devel area. This new package, the successor to survival4, has a more stable likelihood maximiser for parametric survival models and incorporates penalised likelihoods for adding smoothing splines, ridge regression, and (approximately) frailties to survival

I am confused on how the log-likelihood is calculated in a parametric survival problem with frailty. I see a contradiction in the frailty() help file vs. the source code of frailty.gamma(), frailty.gaussian() and frailty.t(). The function frailty.gaussian() appears to calculate the penalty as the negative log-density of independent Gaussian variables, as one would expect: >

Dear all, Has anybody tried numerical solving of ODE's and/or transport equations in R? (Don't ask how we ended up in using R for this job, in the first place!) More precisely, does anybody know any technical issue that could make the work insecure in the sense of propagation of errors? Is there any track of evidence that R is, in this kind of task, less reliable than e.g. MatLab?

On Tue, 6 May 2003, Jerome Asselin wrote: > > I am confused on how the log-likelihood is calculated in a parametric > survival problem with frailty. I see a contradiction in the frailty() help > file vs. the source code of frailty.gamma(), frailty.gaussian() and > frailty.t(). > > The function frailty.gaussian() appears to calculate the penalty as the > negative

Hi Greg, For quasi families I've used extended quasi-likelihood (see Mccullagh and Nelder, Generalized Linear Models 2nd ed, section 9.6) in place of the likelihood/quasi-likelihood in the expression for the (RE)ML score. I hadn't realised that this was possible before the paper was published. best, Simon ps. sorry for slow reply, the original message slipped through my filter for

Dear all, I am trying to use ssanova of the gss package but met some error that I cannot figure out the answer for. Here is the code I am using to explain the problem. library(gss) set.seed(5732) x=(1:100)/100 y=1+3*sin(2*pi*x)+2*(x>0.7)+rnorm(x) x1=rnorm(100) x2=rnorm(100) part.fit=ssanova(y~x, partial=~cbind(x1,x2)) summary(part.fit) part.fit=ssanova(y~x, partial=~cbind(as.numeric(x1),x2))

Hey all, I'm using the gam() function inside the mgcv package to fit a penalised spline to some data. However, I don't quite understand what exactly the intercept it includes by default is / how to interpret it. Ideally I'd like to understand what the intercept is in terms of the B-Spline and/or truncated power series basis representation. Thanks!

hi I'm puzzled as to the relation between the REML score computed by gam and the formula (4) on p.4 here: http://opus.bath.ac.uk/22707/1/Wood_JRSSB_2011_73_1_3.pdf I'm ok with this for poisson, or for quasipoisson when phi=1. However, when phi differs from 1, I'm stuck. #simulate some data library(mgcv) set.seed(1) x1<-runif(500) x2<-rnorm(500)

Dear All, Thank you for the replies to my first thread here: http://r.789695.n4.nabble.com/global-optimisation-with-inequality-constraints-td3799258.html. So far the best result is achieved via a penalised objective function. This was suggested by someone on this list privately. I am still looking into some of the options mentioned in the original thread, but I have been advised that there may

Hi all, I'm using the ssanova function from the gss package to fit smoothing spline anovas, and am running into some difficulty. For my data, I have measurements at 2 milisecond intervals for every observation. Every observation does not have the same duration, so I have scaled the times for each observation to a scale between 0 and 1. I would like to smooth over time, and the following

SEE ALSO ORIGINAL POSTING IN PR#2933 On May 6, 2003 03:58 pm, Thomas Lumley wrote: > > Looking at a wider context in the code > > pfun <- function(coef, theta, ndeath) { > if (theta == 0) > list(recenter = 0, penalty = 0, flag = TRUE) > else { > recenter <- log(mean(exp(coef))) > coef <- coef - recenter