similar to: apply a function to a rolling subset of a vector

Hello: I want to get a rolling estimation of the stdev of my data. Searching the document, I found the function "rollapply" in the zoo package. For example, my series is "c", and i want get a period of 10 days, so i write the command below: roll.sd = rollapply( c, 10, sd, na.pad = TRUE, align = 'right' ) but there is an error in it ,and the computing cannot be

Hi, First, let me thank Jaroslaw for making this survey. I find it quite illuminating. Now the questions: * the #1 solution below (based on cumsum) is numerically unstable. Specifically if you do the runmean on a positive vector you can easily get negative numbers due to rounding errors. Does anyone see a modification which is free of this deficiency? * is it possible to optimize the

I've never seen anything written about multiple comparisons, as in the multcomp package or with TukeyHSD, but using a glm. Do such procedures exist? Are they sensible? Are there any packages in R that implement such comparisons? Thank you. -- Ken Knoblauch Inserm U371 Cerveau et Vision 18 avenue du Doyen Lepine 69675 Bron cedex France Tel: +33 (0)4 72 91 34 77 Fax: +33 (0)4 72 91 34 61

I'm struggling to find a way to add a line (abline, lline or lmline) to a single panel of a lattice plot. In one panel, there is an outlier and I'd like to include a second lmline fit to the data without this point. I've looked through nine pages of the archives on lattice and it wasn't obvious (to me) what additional search keyword to add to hone in on this, although I would

Can anyone explain to me why the AIC values are so different when using glm.nb and glm with a negative.binomial family, from the MASS library? I'm using R 1.8.1 with Mac 0S 10.3.4. >library(MASS) > dfr <- data.frame(c=rnbinom(100,size=2,mu=rep(c(10,20,100,1000),rep(25,4))), + f=factor(rep(seq(1,4),rep(25,4)))) > AIC(nb1 <- glm.nb(c~f, data=dfr)) [1] 1047 >

What does 'with ties in' mean? with some identical elements (par ex., au moins une paire ex-equo) HTH ____________________ Ken Knoblauch Inserm U371, Cerveau et Vision Department of Cognitive Neurosciences 18 avenue du Doyen Lepine 69675 Bron cedex France tel: +33 (0)4 72 91 34 77 fax: +33 (0)4 72 91 34 61 portable: 06 84 10 64 10 http://www.lyon.inserm.fr/371/

Is there a simple function in R to get a moving standard deviation (i.e. for the last x samples)? My goal is to plot bollinger bands around a moving average for price data. I use kernel smoothing for the moving average. cheers and thanks! over and out -- doktora

Hi, I'm trying to calculate the value of the variable, dp, below, in the argument to the integral of dnorm(x-dp) * pnorm(x)^(m-1). This corresponds to the estimate of the sensitivity of an observer in an m-alternative forced choice experiment, given the probability of a correct response, Pc, a Gaussian assumption for the noise and no bias. The function that I wrote below gives me an error:

To add to the wish-list for "convolve": For modeling processes that decay exponentially in time, e.g., fluorescence, it is desirable to have a function that convolves an arbitrary vector with an exponential using an iterative method. In the TIMP package (which won't be on CRAN till R 2.5.0 is official, but is for now at www.nat.vu.nl/~kate/TIMP) we implemented this special-purpose

for the moment, I'm using: .onAttach <- function(libname, pkgname) { .bbg.db.conn <<- dbConnect(dbDriver("PostgreSQL"), user="blah","blah") } .onUnload <- function(libpath) { dbDisconnect(.bbg.db.conn) } which results in a hidden global variable in the global environment. I would prefer to make the assignment only in the package namespace.

I have been experimenting with convolve(). What I know about convolution I learned from engineering (they call this stuff linear systems theory), not statistics, so maybe this is all just a matter of different conventions. BUT I notice very weird things with convolve(). 1. First example, from the classic Bracewell The Fourier transform and its applications, chap 3 (p.32 in 2nd edition): {2 2 3 3

Dear R People: Here is something that I am sure is very simple. I'm just trying to re-create the C convolution example in the Extensions manual. Here is the subroutine: void convolve(double *a, int *na, double *b, int *nb, double *ab) { R_len_t i, j, nab = *na + *nb - 1; for(i = 0; i < nab; i++) ab[i] = 0.0; for(i = 0; i < *na; i++) for(j = 0; j < *nb; j++) ab[i + j] += a[i] *

Dear R users, I want to call C code via the .C or .Call interface. This works fine with integer values but using doubles the array received in C will be set to zeros. I have tried the convolve examples (Writing R extensions, chapter 5.2) and still the resulting array consists of zeros. My code (shortened for my purposes. Original did not work either):

Hi, Lately I run into a problem that my code R code is spending hours performing simple moving window statistical operations. As a result I did searched archives for alternative (faster) ways of performing: mean, max, median and mad operation over moving window (size 81) on a vector with about 30K points. And performed some timing for several ways that were suggested, and few ways I come up

We obtained some disturbing results from convolve() (inaccuracies and negative probabilities). We'll try to make the context clear in as few lines as possible... Our function panjer() (code below) basically computes recursively the probability mass function of a compound Poisson distribution. When the Poisson parameter lambda is very large, the starting value of the recursive scheme ---

Hi, I compiled a C program file on Borland C++ 5.5 compiler to get one dll output (as instructed in the file readme.package). The C program file is just the example on page 31 of "writing R extensions": void convolve(double *a, int *na, double *b, int *nb, double *ab) { int i, j, nab = *na + *nb - 1; for(i = 0; i < nab; i++) ab[i] = 0.0; for(i = 0; i < *na; i++) for(j = 0; j

Hi, Man page for 'convolve' says: conj: logical; if 'TRUE', take the complex _conjugate_ before back-transforming (default, and used for usual convolution). The complex conjugate of 'x', of 'y', of both? In fact it seems that it takes the complex conjugate of 'y' only which is OK but might be worth mentioning because (1) conj=TRUE is the

Have you noticed any problems with big dates (>=1/1/2040) in R? Here is the bit of code that I'm having trouble with: > test.date <- strptime("1/1/2040",format="%m/%d/%Y") > > unlist(test.date) sec min hour mday mon year wday yday isdst 0 0 0 1 0 140 0 0 0 > > date.plus.one <- as.POSIXct(test.date) +

Hi, the function stats::convolve does not mention efficient usage of the underlying FFT algorithm, such as (a) if type="circular", then length(x)=length(y) should have many factors (e.g. length(x) = length(y) = 2^n) (b) if type="open" or "filter", then length(x)+length(y)-1 should have many factors (e.g. length(x)+length(y)-1 = 2^n) In particular the latter may

I see that plot.default uses deparse(substitute(x)) to extract the character name of an argument and put it on the vertical axis. Hence: foo <- 1:10 plot( foo ) will put the label "foo" on the vertical axis. However, for a function that takes a "..." list as an input, I can only extract the first argument name: x <- 1:10 y <- 10:20 foo <- function(...) {