similar to: Savitzky-Golay smoothing -- an R implementation

Hi I am running a Savitzky-Golay smoothing filter (http://tolstoy.newcastle.edu.au/R/help/04/02/0385.html) for variables in my dataset, dim (272:90). I managed to run the code for individual variables in the dataset and then combine the results into a single dataset. My novice attempt at this task is shown below csg<-NULL for (i in 1:ncol(data.all)) {

If I'm not mistaken, that's sort of local polynomial with even degree and fixed bandwidth (based on my own interpretation of description in Numerical Recipes). You can do that with functions in the KernSmooth package. HTH, Andy > -----Original Message----- > From: wolf at micro-biolytics.com [mailto:wolf at micro-biolytics.com] > Sent: Friday, March 21, 2003 1:43 PM > To:

Hi When we use the sav_gol command in R , it shows an error which says: " error in as.matrix". We've downloaded the necessary packages. Kindly help us with this issue. If there is any other function to perform Savitzky-Golay smoothing in R, please let me know. With Regards Reynolds

I got a question from a fellow PhD student that work with spectrum analysis in Excel and now he has lots of spectrums that needs to be smoothed, which would be nice to be able to do in batch. Is there an R package that can do: Savitzky-Golay smoothing for reflectance spectral data or a function that does something similar. _______________________________________ Henrik Andersson

Dear all, Sorry if this is slightly off the track as far as R is concerned, but I have been using the Savitzky-Golay filter to estimate some derivatives of interest. I am wondering however, if anyone has seen anything in the literature (or has any ideas) of how these estimates perform asymptotically. Does anyone know what the rate of convergence is for these? Thanks, matt.

Hi, Does anyone know how to use weights and generate error bounds for Savitzky-Golay? I have a (smallish) set of points y equally spaced each with a known error and would like to smooth them using S-G but so as to take into account the error already have and construct new error bounds around them that take into account the errors they had at the beginning and the erros they get as a result

Dear all, Sorry if this is slightly off the track as far as R is concerned, but I have been using the Savitzky-Golay filter to estimate some derivatives of interest. I am wondering however, if anyone has seen anything in the literature (or has any ideas) of how these estimates perform asymptotically. Does anyone know what the rate of convergence is for these as the sample size increases?

Dear All, I have a pair of .cpp and .def file can be compiled using VC++ and works perfectly well in S-PLUS. I wanted to do the same for R; so I followed the guidline given in "Building from a source-code library under Windows" as much as possible and manage to compile them using VC++ and call it from R. But it gives different answer from the one called from S-Plus. I know that I did

Dear all; I'm kind of confused with the results obtained using the ginv function from package MASS and pinv function from Matlab. Accroding to the documentation both functions performs a Moore-Penrose generalized inverse of a matrix X. The problem is when I change the tolerance value, say to 1E-3. Here is some output from ginv 195.2674402 235.6758714 335.0830253 8.977515484 -291.7798965

Dear alltogether, I tried pan() to impute NAs for longitudinal data. The terminology in the following output follows the pan manpage. No data are attached to this script as this may be too huge. y = 15 responses pred = at first just intercept was tried (later on covariates should follow) subj = 168 different subjects with 4 to 6 observations for each subject at time points t1, t2, ..., t6 #

Hi, I am writing some basic smoothers in R for cleaning some spectral data. I wanted to see if I could get close to matlab for speed, so I was trying to compare SparseM with Matrix to see which could do the choleski decomposition the fastest. Here is the function using SparseM difsm <- function(y, lambda, d){ # Smoothing with a finite difference penalty # y: signal to be smoothed #

Hi the list. I do not understand the philosophy behind numeric and integer. - 1 is numeric (which I find surprising) - 2 is numeric. - 1:2 is integer. Why is that ? Christophe

Hello all. I am trying to run PAN, multilevel multiple imputation program, in R to impute missing data in a longitudinal dataset. I could successfully run the multiple imputation when I only imputed one variable. However, when I tried to impute a time-varying covariate as well as a response variable, I received an error message, “Error: subscript out of bounds.” Can anyone tell if my commands

Dear useRs, This is the first time I post to this list and I would appreciate any help available. I've used the excellent mgcv package for a while now to investigate geographical patterns of language variation, and it has has always worked without any problems for me. The problem below occurs using R 2.14.0 (both 32 and 64 bit versions in Windows and the 64 bit version in Unix) and mgcv (both

Apologies if this is this too obscure for R-help. In package splines, ns(x,,knots,intercept=TRUE) produces an n by K+2 matrix N, the values of K+2 basis functions for the natural splines with K (internal) knots, evaluated at x. It does this by first generating an n by K+4 matrix B of unconstrained splines, then postmultiplying B by H, a K+4 by K+2 representation of the nullspace of C (2 by K+4),

Hello all, I am working on a longitudinal study of children in the UK and trying the PAN package for imputation of missing data, since it fulfils the critical criteria of taking into account individual subject trend over time as well as population trend over time. In order to validate the procedure I have started by deleting some known values ?we have 6 annual measures of height on 300 children

Hi, I've tried to run the multiple imputation for longitudinal data using Pan in R. The trimmed data consist 10 individuals across 4 time points. Following the example in panex.R, I imputated the model with one y variable and only random intercept. It worked well. The next step for me is to imputate a model with one y variable and both random intercept and slope. The program ran well, but

Dear R-List, I apologize for not posting a reproducible example - the reason is that I actually do not succeed in reproducing my specific problem with generated data. Let me still describe the problem: I want to impute missing data using the "pan" package. a) It works, when I use a fraction of my data (e.g. 200 out of 44000 cases) b) It works, when I generate a dataset of equal

On examining non-linearity of Cox coefficients with penalized splines - I have not been able to dig up a completely clear description of the test performed in R or S-plus. >From the Therneau and Grambsch book (2000 - page 126) I gather that the test reported for "linear" has as its null hypothesis that the spline coefficient is the same at the center of basis. Thus, in the example

OK, I searched through the archives and didn't find anything so I must be doing something stupid. I just compiled 2.6.3 and am trying to get it to run on linux 2.6.8.1 kernel with a more or less fedora core2 environment. I was trying between two systems but I have narrowed it down to what I believe is a pretty simple case. Here is my setup/test: Last login: Mon Oct 4 19:30:49 2004