Re: [R] problems saving and loading (PLMset) objects
Erm, Jim I am loading in the affyPLM package first (when needed) and this was a question based on loading/saving R objects. PLMset was an example. Many thanks, Quin -Original Message- From: James W. MacDonald [mailto:[EMAIL PROTECTED] Sent: 31 July 2007 14:54 To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] problems saving and loading (PLMset) objects Hi Quin, First off, you should ask questions about Bioconductor packages on the BioC listserv rather than R-help. Anyway, I don't think your PLMset objects are coming out all wrong - it doesn't appear that you are loading the affyPLM package first, which is required for R to know anything about the PLMset object (this object is defined in affyPLM, so without the package R has no idea what it is). Best, Jim Quin Wills wrote: Hi I'm running the latest R on a presumably up to date Linux server. 'Doing something silly I'm sure, but can't see why my saved PLMset objects come out all wrong. To use an example: Setting up an example PLMset (I have the same problem no matter what example I use) library(affyPLM) data(Dilution) # affybatch object Dilution = updateObject(Dilution) options(width=36) expr - fitPLM(Dilution) This works, and I'm able to get the probeset coefficients with coefs(expr). until I save and try reloading: save(expr, file=expr.RData) rm(expr) # just to be sure expr - load(expr.RData) Now, running coefs(expr) says: Error in function (classes, fdef, mtable) : unable to find an inherited method for function coefs, for signature character Trying str(exp) just gives the following: chr exp expr.Rdata appears to save properly (in that there is an actual file with notable size in my working directory). Thanks in advance, Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- James W. MacDonald, M.S. Biostatistician Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 734-647-5623 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] problems saving and loading (PLMset) objects
Hi I'm running the latest R on a presumably up to date Linux server. 'Doing something silly I'm sure, but can't see why my saved PLMset objects come out all wrong. To use an example: Setting up an example PLMset (I have the same problem no matter what example I use) library(affyPLM) data(Dilution) # affybatch object Dilution = updateObject(Dilution) options(width=36) expr - fitPLM(Dilution) This works, and I'm able to get the probeset coefficients with coefs(expr). until I save and try reloading: save(expr, file=expr.RData) rm(expr) # just to be sure expr - load(expr.RData) Now, running coefs(expr) says: Error in function (classes, fdef, mtable) : unable to find an inherited method for function coefs, for signature character Trying str(exp) just gives the following: chr exp expr.Rdata appears to save properly (in that there is an actual file with notable size in my working directory). Thanks in advance, Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] problems saving and loading (PLMset) objects
Ah, didn't realize that I couldn't re-assign in one step... I was trying to load in various data (on separate occasions), using a common object name to run through some template code. Many thanks! -Original Message- From: jim holtman [mailto:[EMAIL PROTECTED] Sent: 30 July 2007 23:51 To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] problems saving and loading (PLMset) objects you just need to say: load(expr.RData) You should not be assigning it to 'expr' since it is already 'load'ed On 7/30/07, Quin Wills [EMAIL PROTECTED] wrote: Hi I'm running the latest R on a presumably up to date Linux server. 'Doing something silly I'm sure, but can't see why my saved PLMset objects come out all wrong. To use an example: Setting up an example PLMset (I have the same problem no matter what example I use) library(affyPLM) data(Dilution) # affybatch object Dilution = updateObject(Dilution) options(width=36) expr - fitPLM(Dilution) This works, and I'm able to get the probeset coefficients with coefs(expr). until I save and try reloading: save(expr, file=expr.RData) rm(expr) # just to be sure expr - load(expr.RData) Now, running coefs(expr) says: Error in function (classes, fdef, mtable) : unable to find an inherited method for function coefs, for signature character Trying str(exp) just gives the following: chr exp expr.Rdata appears to save properly (in that there is an actual file with notable size in my working directory). Thanks in advance, Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem you are trying to solve? __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] multiple rugs on a single plot
Hi I could only find some discussion on this wrt lattice graphics (which I'm not using). Apologies if I'm missing something obvious. I'd like to produce 3 rug plots under a kernel density plot for a population. The population is subdivided into 3 subpopulations, which I'd like the rug plots to highlight. Naturally, when I do 3 rug plots, they all plot over each other. I'd like 3 parallel rug plots along the x-axis. But how. Thanks in advance, Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] multiple rugs on a single plot
Excellent! Thanks. -Original Message- From: Dimitris Rizopoulos [mailto:[EMAIL PROTECTED] Sent: 17 July 2007 13:40 To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] multiple rugs on a single plot you could use different colours, e.g., x1 - rnorm(100, -2.5, 1) x2 - rnorm(100, 0, 1) x3 - rnorm(100, 2.5, 1) x - c(x1, x2, x3) plot(density(x)) rug(x1, col = red) rug(x2, col = black) rug(x3, col = blue) or something like the following: plot(density(x)) len - 0.005 ds - 0.001 segments(x1, -1, x1, 0) segments(x2, 0 + ds, x2, len) segments(x3, len + ds, x3, 2*len) I hope it helps. Best, Dimitris Dimitris Rizopoulos Ph.D. Student Biostatistical Centre School of Public Health Catholic University of Leuven Address: Kapucijnenvoer 35, Leuven, Belgium Tel: +32/(0)16/336899 Fax: +32/(0)16/337015 Web: http://med.kuleuven.be/biostat/ http://www.student.kuleuven.be/~m0390867/dimitris.htm - Original Message - From: Quin Wills [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Sent: Tuesday, July 17, 2007 12:50 PM Subject: [R] multiple rugs on a single plot Hi I could only find some discussion on this wrt lattice graphics (which I'm not using). Apologies if I'm missing something obvious. I'd like to produce 3 rug plots under a kernel density plot for a population. The population is subdivided into 3 subpopulations, which I'd like the rug plots to highlight. Naturally, when I do 3 rug plots, they all plot over each other. I'd like 3 parallel rug plots along the x-axis. But how. Thanks in advance, Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] PCA with not non-negative definite covariance
Thank you... I will definitely check that up. Quin -Original Message- From: Stéphane Dray [mailto:[EMAIL PROTECTED] Sent: 27 July 2006 09:04 AM To: Quin Wills Cc: 'Berton Gunter'; r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance As said by Pierre Bady, an answer to your question is NIPALS analysis. PCA is usually obtained by the diagonalization of a variance-covariance matrix. But it can also be obtained by an iterative proedure which consists in two regressions. NIPLAS is an implementation of this iterative procedure and is strictly equivalent to PCA when there is no missing values. The adavantage of NIPALS is that it can be used with missing values. However, note that the convergence is not always obtained (it depends of the number and distribution of missing values). You can find a description of the method and the algorithm here: http://biomserv.univ-lyon1.fr/~dray/articles/SD165.html Sincerely, Quin Wills wrote: My apologies (in response to the last 2 replies). I should write sensibly - including subject titles that make grammatical sense. (1) By analogous, I mean that using classical MDS with Euclidian distance is equivalent to plotting the first k principle components. (2) Agreed re. distribution assumptions. (3) Agreed re. the need to use some kind of imputation for calculating distances. I'm thinking pairwise exclusion for correlation. Re. why I want to do this is simply for graphically representing my data. Quin -Original Message- From: Berton Gunter [mailto:[EMAIL PROTECTED] Sent: 26 July 2006 05:10 PM To: 'Quin Wills'; [EMAIL PROTECTED] Cc: r-help@stat.math.ethz.ch Subject: RE: [R] PCA with not non-negative definite covariance Not sure what completely analagous means; mds is nonlinear, PCA is linear. In any case, the bottom line is that if you have high dimensional data with many missing values, you cannot know what the multivariate distribution looks like -- and you need a **lot** of data with many variables to usefully characterize it anyway. So you must either make some assumptions about what the distribution could be (including imputation methodology) or use any of the many exploratory techniques available to learn what you can. Thermodynamics holds -- you can't get something for nothing (you can't fool Mother Nature). -- Bert Gunter Genentech Non-Clinical Statistics South San Francisco, CA The business of the statistician is to catalyze the scientific learning process. - George E. P. Box -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Quin Wills Sent: Wednesday, July 26, 2006 8:44 AM To: [EMAIL PROTECTED] Cc: r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance Thanks. I suppose that another option could be just to use classical multi-dimensional scaling. By my understanding this is (if based on Euclidian measure) completely analogous to PCA, and because it's based explicitly on distances, I could easily exclude the variables with NA's on a pairwise basis when calculating the distances. Quin -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: 25 July 2006 09:24 AM To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance Hi , hi all, Am I correct to understand from the previous discussions on this topic (a few years back) that if I have a matrix with missing values my PCA options seem dismal if: (1) I dont want to impute the missing values. (2) I dont want to completely remove cases with missing values. (3) I do cov() with use=pairwise.complete.obs, as this produces negative eigenvalues (which it has in my case!). (4) Maybe you can use the Non-linear Iterative Partial Least Squares (NIPALS) algorithm (intensively used in chemometry). S. Dray proposes a version of this procedure at http://pbil.univ-lyon1.fr/R/additifs.html. Hope this help :) Pierre -- Ce message a été envoyé depuis le webmail IMP (Internet Messaging Program) -- No virus found in this incoming message. -- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- Stéphane DRAY ([EMAIL PROTECTED] ) Laboratoire BBE-CNRS-UMR-5558, Univ. C. Bernard - Lyon I 43, Bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France Tel: 33 4 72 43 27 57 Fax: 33 4 72 43 13 88 http://biomserv.univ-lyon1.fr/~dray/ -- No virus found in this incoming message. -- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help
Re: [R] PCA with not non-negative definite covariance
Thanks. I suppose that another option could be just to use classical multi-dimensional scaling. By my understanding this is (if based on Euclidian measure) completely analogous to PCA, and because it's based explicitly on distances, I could easily exclude the variables with NA's on a pairwise basis when calculating the distances. Quin -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: 25 July 2006 09:24 AM To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance Hi , hi all, Am I correct to understand from the previous discussions on this topic (a few years back) that if I have a matrix with missing values my PCA options seem dismal if: (1) I dont want to impute the missing values. (2) I dont want to completely remove cases with missing values. (3) I do cov() with use=pairwise.complete.obs, as this produces negative eigenvalues (which it has in my case!). (4) Maybe you can use the Non-linear Iterative Partial Least Squares (NIPALS) algorithm (intensively used in chemometry). S. Dray proposes a version of this procedure at http://pbil.univ-lyon1.fr/R/additifs.html. Hope this help :) Pierre -- Ce message a été envoyé depuis le webmail IMP (Internet Messaging Program) -- No virus found in this incoming message. -- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] PCA with not non-negative definite covariance
My apologies (in response to the last 2 replies). I should write sensibly - including subject titles that make grammatical sense. (1) By analogous, I mean that using classical MDS with Euclidian distance is equivalent to plotting the first k principle components. (2) Agreed re. distribution assumptions. (3) Agreed re. the need to use some kind of imputation for calculating distances. I'm thinking pairwise exclusion for correlation. Re. why I want to do this is simply for graphically representing my data. Quin -Original Message- From: Berton Gunter [mailto:[EMAIL PROTECTED] Sent: 26 July 2006 05:10 PM To: 'Quin Wills'; [EMAIL PROTECTED] Cc: r-help@stat.math.ethz.ch Subject: RE: [R] PCA with not non-negative definite covariance Not sure what completely analagous means; mds is nonlinear, PCA is linear. In any case, the bottom line is that if you have high dimensional data with many missing values, you cannot know what the multivariate distribution looks like -- and you need a **lot** of data with many variables to usefully characterize it anyway. So you must either make some assumptions about what the distribution could be (including imputation methodology) or use any of the many exploratory techniques available to learn what you can. Thermodynamics holds -- you can't get something for nothing (you can't fool Mother Nature). -- Bert Gunter Genentech Non-Clinical Statistics South San Francisco, CA The business of the statistician is to catalyze the scientific learning process. - George E. P. Box -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Quin Wills Sent: Wednesday, July 26, 2006 8:44 AM To: [EMAIL PROTECTED] Cc: r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance Thanks. I suppose that another option could be just to use classical multi-dimensional scaling. By my understanding this is (if based on Euclidian measure) completely analogous to PCA, and because it's based explicitly on distances, I could easily exclude the variables with NA's on a pairwise basis when calculating the distances. Quin -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: 25 July 2006 09:24 AM To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] PCA with not non-negative definite covariance Hi , hi all, Am I correct to understand from the previous discussions on this topic (a few years back) that if I have a matrix with missing values my PCA options seem dismal if: (1) I dont want to impute the missing values. (2) I dont want to completely remove cases with missing values. (3) I do cov() with use=pairwise.complete.obs, as this produces negative eigenvalues (which it has in my case!). (4) Maybe you can use the Non-linear Iterative Partial Least Squares (NIPALS) algorithm (intensively used in chemometry). S. Dray proposes a version of this procedure at http://pbil.univ-lyon1.fr/R/additifs.html. Hope this help :) Pierre -- Ce message a été envoyé depuis le webmail IMP (Internet Messaging Program) -- No virus found in this incoming message. -- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. -- No virus found in this incoming message. -- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] PCA with not non-negative definite covariance
Am I correct to understand from the previous discussions on this topic (a few years back) that if I have a matrix with missing values my PCA options seem dismal if: (1) I dont want to impute the missing values. (2) I dont want to completely remove cases with missing values. (3) I do cov() with use=pairwise.complete.obs, as this produces negative eigenvalues (which it has in my case!). This seems like such a shame as I would like to use PCA to plot my clustering results. Any wisdom? Quin -- Checked by AVG Free Edition. [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] How best to deal with returned errors?
Hi, What is the best general strategy to prevent returned errors from interrupting whatever it is I am running? Only by using options()? This is a problem for me in 2 particular cases: (i) An error breaking my loops. (ii) I would like to run some regressions where it automatically will try a specified number of increasingly robust options until the regression doesnt fail. Using try(), especially for the second case, seems a bit clunky for my needs. As Ive never really automated the handling of my errors I was wondering if there is perhaps a bit of simple wisdom or pointers in the right direction on the matter. Thanks, Quin -- Checked by AVG Free Edition. [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Weighting cluster variables in R?
Are there functions to weight variables for clustering in R? I can't seem to find anything, so apologies if there is. I am particularly interested in weighting variables (starting with kmeans) to optimise inter/intra-cluster distances. It seems to me that if certain variables do show a strong cluster structure, this would be a wise thing to do. Any advice welcome. Quin [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] How to use a validation set rather than the default cross-validation in rpart() ?
Many thanks. I'm using it for pruning and was hoping that rpart allows use of a validation set rather than cross-validation for generating a CP/error table. -Original Message- From: Uwe Ligges [mailto:[EMAIL PROTECTED] Sent: 03 May 2006 07:53 To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] How to use a validation set rather than the default cross-validation in rpart() ? Quin Wills wrote: I want use a validation set for my classification tree rather than the default 10-fold validation in rpart() but can't see which arguments to use to get this right. Advice appreciated thanks. I assume that this is possible! You cannot for the internal algorithm that optimizes the splits of the tree. Of course you can do so for estimating the misclassification rate (or whatever), but this has nothing to do with rpart() itself Uwe Ligges __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] How to use a validation set rather than the default cross-validation in rpart() ?
Is it not true that cross-validation can sometimes over estimate classification error - versus bringing in an external validation data set and checking its classification error? I was trying to test this out, but from what I see either way seems to be much of muchness. -Original Message- From: Prof Brian Ripley [mailto:[EMAIL PROTECTED] Sent: 03 May 2006 10:33 To: Quin Wills Cc: 'Uwe Ligges'; r-help@stat.math.ethz.ch Subject: Re: [R] How to use a validation set rather than the default cross-validation in rpart() ? On Wed, 3 May 2006, Quin Wills wrote: Many thanks. I'm using it for pruning and was hoping that rpart allows use of a validation set rather than cross-validation for generating a CP/error table. Since it is not documented how to, why do you expect to? Indeed, why do you think it would be a good idea? -Original Message- From: Uwe Ligges [mailto:[EMAIL PROTECTED] Sent: 03 May 2006 07:53 To: Quin Wills Cc: r-help@stat.math.ethz.ch Subject: Re: [R] How to use a validation set rather than the default cross-validation in rpart() ? Quin Wills wrote: I want use a validation set for my classification tree rather than the default 10-fold validation in rpart() but can't see which arguments to use to get this right. Advice appreciated thanks. I assume that this is possible! You cannot for the internal algorithm that optimizes the splits of the tree. Of course you can do so for estimating the misclassification rate (or whatever), but this has nothing to do with rpart() itself Uwe Ligges __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Brian D. Ripley, [EMAIL PROTECTED] Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UKFax: +44 1865 272595 __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] How to use a validation set rather than the default cross-validation in rpart() ?
I want use a validation set for my classification tree rather than the default 10-fold validation in rpart() but can't see which arguments to use to get this right. Advice appreciated thanks. I assume that this is possible! [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] How to get around heteroscedasticity with non-linear least squares in R?
I am using nls to fit dose-response curves but am not sure how to approach more robust regression in R to get around the problem of the my error showing increased variance with increasing dose. My understanding is that rlm or lqs would not be a good idea here. 'Fairly new to regression work, so apologies if I'm missing something obvious. [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] How to get around heteroscedasticity with non-linear leas t squares in R?
Thank you all, this has been a great help (including the methodological
advice). Very interesting - I'll be sure to read the lecture.
Quin
-Original Message-
From: Liaw, Andy [mailto:[EMAIL PROTECTED]
Sent: 22 February 2006 01:18
To: 'Brian S Cade'; [EMAIL PROTECTED]
Cc: Quin Wills; r-help@stat.math.ethz.ch; [EMAIL PROTECTED]
Subject: RE: [R] How to get around heteroscedasticity with non-linear leas t
squares in R?
From: Brian S Cade
Instead of thinking that the heteroscedasticity is a nuisance and
something to get around, i.e, just wanting weighted
estimates of the
mean function, you might want to think about what
heteroscedasticity is
telling you and estimate some other quantities.
Indeed! See Prof. Carroll's 2002 Fisher Lecture:
http://www.stat.tamu.edu/ftp/pub/rjcarroll/2003.papers.directory/published_F
isher_Lecture.pdf
(There's Powerpoint file on his web page, too.)
Andy
Heteroscedasticity is
telling you that the conditional distributions don't change
at a constant
rate across all portions of the distribution (think
percentiles or more
generally quantiles) and, therefore, a function for the mean
(no matter
how precisely estimated) cannot tell you all there is to know
about your
dose-response relation. Why not go after estimating the conditional
quantile functions directly with nonlinear quantile
regression, function
nlrq() in the quantreg package?
Brian
Brian S. Cade
U. S. Geological Survey
Fort Collins Science Center
2150 Centre Ave., Bldg. C
Fort Collins, CO 80526-8818
email: [EMAIL PROTECTED]
tel: 970 226-9326
Kjetil Brinchmann Halvorsen [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
02/21/2006 03:31 PM
Please respond to
[EMAIL PROTECTED]
To
Quin Wills [EMAIL PROTECTED]
cc
r-help@stat.math.ethz.ch
Subject
Re: [R] How to get around heteroscedasticity with non-linear
least squares
in R?
Quin Wills wrote:
I am using nls to fit dose-response curves but am not sure how to
approach
more robust regression in R to get around the problem of
the my error
showing increased variance with increasing dose.
package sfsmisc has rnls (robust nls)
which might be of use.
Kjetil
My understanding is that rlm or lqs would not be a good
idea here.
'Fairly new to regression work, so apologies if I'm missing
something
obvious.
[[alternative HTML version deleted]]
__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
[[alternative HTML version deleted]]
__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
--
Notice: This e-mail message, together with any attachments,...{{dropped}}
__
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] How to access values returned by R functions (to put into vectors)?
The question is general for all functions, but here is a specific example - # For the logistic regression of the following correlated variables: C - c(457, 1371, 4113, 12339, 37017, 111051, 333153, 999459) E - c(0.003858377, 0.014334578, 0.014092836, 0.737950754, 0.996371828, 0.997482379, 1.005569257, 0.994382856) # The nls function: A = nls(E~(Em*C^p)/(C50^p + C^p), start = list(Em=0.8, p=3, C50=1e3)) # Returns the following parameter estimates for Em, p and C50: Nonlinear regression model model: E ~ (Em * C^p)/(C50^p + C^p) data: parent.frame() Emp C50 0.99891134.7957189 9934.6481397 residual sum-of-squares: 0.0002856567 How do I access these parameter values from this output/function so that it would go into a vector c(Em,p,C50)? [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Installing SJava (I'm almost there, just a little more help please!....please!)
Hi. Day three and Im still struggling with this. Any advice to overcome the final hurdle will be enormously appreciated. I now have all the right Java applications etc. in their right places and have managed to get rid of most errors but still get this: Making package SJava Building JNI header files... adding build stamp to DESCRIPTION running src/Makefile.win (cd .. ; ./configure.win c:/PROGR~1/R/rw200l) /configure.win: not found make[3]: *** [config] Error 127 make[2]: *** [srcDynLib] Error 2 make[1]: *** [all] Error 2 Make: *** [pk9SJava] Error 2 *** Installation of SJava failed *** Removing c:/PROGR~1/R/rw200l/library/SJava I am Windows XP with SJava (SJava_0.68-0.tar.gz) downloaded to my c drive (c:\SJava_0.68-0.tar.gz). R is rw2001 (c:\Program Files\R\rw2001). I am using the following R CMD INSTALL c:\SJava-0.68-0.tar.gz. Why is the configure.win file not being found? Where is it looking for it? My eternal gratitude to anybody willing to take me out of my pain. --- [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] Installing SJava (I'm almost there, just a little more help please!....please!)
Hi Duncan Thank you for responding... I apologise for being so ignorant. I presume that is a UNIX command - so have just downloaded cygwin (read about that today) and ran your suggested line. I get: Chmod: cannot acess 'configure.win'. No such file or directory I am assuming that isn't good. Is this the configure file I find in SJava? Do I need to move it somewhere for this to work? All of the best, Quin -Original Message- From: Duncan Temple Lang [mailto:[EMAIL PROTECTED] Sent: 27 July 2005 09:30 PM To: Quin Wills Subject: Re: [R] Installing SJava (I'm almost there, just a little more help please!please!) Hi. It has been a long time since I looked at the Windows side of things of SJava. Is configure.win present _AND_ executable. Make certain that it is by using chmod +x configure.win D. Quin Wills wrote: Hi. Day three and I’m still struggling with this. Any advice to overcome the final hurdle will be enormously appreciated. I now have all the right Java applications etc. in their right places and have managed to get rid of most errors but still get this: —Making package SJava — Building JNI header files... adding build stamp to DESCRIPTION running src/Makefile.win (cd .. ; ./configure.win c:/PROGR~1/R/rw200l) /configure.win: not found make[3]: *** [config] Error 127 make[2]: *** [srcDynLib] Error 2 make[1]: *** [all] Error 2 Make: *** [pk9—SJava] Error 2 *** Installation of SJava failed *** Removing ‘c:/PROGR~1/R/rw200l/library/SJava’ I am Windows XP with SJava (SJava_0.68-0.tar.gz) downloaded to my c drive (c:\SJava_0.68-0.tar.gz). R is rw2001 (c:\Program Files\R\rw2001). I am using the following “R CMD INSTALL c:\SJava-0.68-0.tar.gz”. Why is the configure.win file not being found? Where is it looking for it? My eternal gratitude to anybody willing to take me out of my pain. --- [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html -- Duncan Temple Lang[EMAIL PROTECTED] Department of Statistics work: (530) 752-4782 371 Kerr Hall fax: (530) 752-7099 One Shields Ave. University of California at Davis Davis, CA 95616, USA --- Incoming mail is certified Virus Free. --- __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Anybody have a binary version of SJava for rw2001 (Windows)?
I am not a techie and have been struggling 2 days solid to try and install SJava (the source from http://www.omegahat.org/RSJava/). Does anybody have a binary file for me (I am Windows XP and rw2001)? I have tried installing Perl, mingwin and the cygwin tools but still no luck. When I try R CMD INSTALL c:\SJava_0.78-0.tar.gz I get the following (and havent a clue what it could mean): -Making package SJava-- Building JNI header files... Extracting the classes from Environment.jar /jdk1.3/bin/jar: not found RForeignReference /jdk1.3/bin/javah: not found ROmegahat Interpreter /jdkl.3/bin/javah: not found REvaluator /jdkl.3/bin/Javah: not found RManualFunctionActionListener /jdk1.3/bin/javah: not found /jdkl.3/bin/javah: not found adding build stamp to DESCRIPTION running src/Makefile.win (cd .. ; ./configure.win c:/PROGRW1/R/rw200l) /configure.win: not found make[3]: *** [conf ig] Error 127 make[2]: *** [srcDynLib] Error 2 make[1]: *** [all] Error 2 make: *** [pkgSJava] Error 2 *** Installation of SJava failed *** --- [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Installing SJava
Apologies, I am a non-techie so am finding installing SJava (which I need for RMAGEML) very frustrating indeed. I am running rw2001 on WindowsXP with Java - jre1.5.0_02. Could somebody please explain to me step-by-step how to install it as I have tried all the help files. The websites (HYPERLINK http://www.omegahat.org/RSJava/http://www.omegahat.org/RSJava/) instructions for installing the Windows source are definitely not Windows commands (are they UNIX? The file is not even a zip as stated in the instructions!). I have also tried using the command R CMD INSTALL HYPERLINK http://www.omegahat.org/RSJava/SJava_0.68-0.tar.gzSJava_0.68-0.tar.gz in R but just get a syntax error. I have tried changing the file to a zip and then installing from the R GUI (install packages from local zip) but when I type library(SJava) I get Error in library(SJava) : 'SJava' is not a valid package -- installed 2.0.0?. Please could somebody explain any way to do this (step by step) to me like Im a two year old. I have no interest in actually calling Java, I just want to be able to read MAGE-ML. I have saved the source file (HYPERLINK http://www.omegahat.org/RSJava/SJava_0.68-0.tar.gzSJava_0.68-0.tar.gz) in my C drive (viz. c:\ HYPERLINK http://www.omegahat.org/RSJava/SJava_0.68-0.tar.gzSJava_0.68-0.tar.gz) wh at now? --- [[alternative HTML version deleted]] __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
