[R] VGAM package : Frechet distribution - 2 parameter estimation
Dear R forum, I am trying to execute following code (Page no 259 - VGAM.pdf) # . library(VGAM) set.seed(123) fdata - data.frame(y1 = rfrechet(nn - 1000, shape = 2 + exp(1))) with(fdata, hist(y1)) fit2 - vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) # . However, I receive following error Error in vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) : object 'frechet' not found Earlier there used to be a function called frechet3 which I guess has been withdrawn by VGAM. Kindly guide Katherine [[alternative HTML version deleted]] __ R-help@r-project.org 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] limit of cmdscale function
Hi
We have a few questions regarding the use of the isoMDS function.
When we run isoMDS function using 60,000 x 60,000 data matrix,
we get the following error message:
cmdscale(d, k) : invalid value of 'n'
Calls: isoMDS - cmdscale
We checked the source code of cmdscale and found the following limitation:
## we need to handle nxn internally in dblcen
if(is.na(n) || n 46340) stop(invalid value of 'n')
1. This cmdscale limitation ('n 46340') is due to the limitation of BLAS and
LAPACK variables(int4) which can only handle '2^31-1' amount of data?
2. Is there any workaround to run isoMDS using large data (i.e. greater than
46340)?
We would like to run isoMDS using a maximum of 150,000x150,000 data matrix.
Best regards
Masayuki Kawashima
Email: kawasim...@jp.fujitsu.com
__
R-help@r-project.org 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] Time dependant covariate
Hello, Can I get help? I have the attached kind of material and I should create a time-dependant covariate from the variate Optime? I have tried using survsplit function but I don't know exactly how to do it. Or I can't get it to work. How can I split every id's follow-up time into two rows according the time of optime (The time they get treatment)? Best regards, Paula Iso-Markku __ R-help@r-project.org 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] VGAM package : Frechet distribution - 2 parameter estimation
On 06/11/2014 06:04, Katherine Gobin wrote: Dear R forum, I am trying to execute following code (Page no 259 - VGAM.pdf) # . library(VGAM) set.seed(123) fdata - data.frame(y1 = rfrechet(nn - 1000, shape = 2 + exp(1))) with(fdata, hist(y1)) fit2 - vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) # . Is it not called frechet2? However, I receive following error Error in vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) : object 'frechet' not found Earlier there used to be a function called frechet3 which I guess has been withdrawn by VGAM. Kindly guide Katherine [[alternative HTML version deleted]] __ R-help@r-project.org 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 message. Checked by AVG - www.avg.com Version: 2015.0.5557 / Virus Database: 4189/8518 - Release Date: 11/05/14 -- Michael http://www.dewey.myzen.co.uk __ R-help@r-project.org 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] VGAM package : Frechet distribution - 2 parameter estimation
Dear Mr Michael, Thanks a lot for your guidance. The pdf file describing VGAM package has mentioned 'frechet' in the example, so I got the error. Regards Katherine On Thursday, 6 November 2014 2:54 PM, Michael Dewey i...@aghmed.fsnet.co.uk wrote: On 06/11/2014 06:04, Katherine Gobin wrote: Dear R forum, I am trying to execute following code (Page no 259 - VGAM.pdf) # . library(VGAM) set.seed(123) fdata - data.frame(y1 = rfrechet(nn - 1000, shape = 2 + exp(1))) with(fdata, hist(y1)) fit2 - vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) # . Is it not called frechet2? However, I receive following error Error in vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) : object 'frechet' not found Earlier there used to be a function called frechet3 which I guess has been withdrawn by VGAM. Kindly guide Katherine [[alternative HTML version deleted]] __ R-help@r-project.org 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 message. Checked by AVG - www.avg.com Version: 2015.0.5557 / Virus Database: 4189/8518 - Release Date: 11/05/14 -- Michael http://www.dewey.myzen.co.uk [[alternative HTML version deleted]] __ R-help@r-project.org 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] VGAM package : Frechet distribution - 2 parameter estimation
Dear Ms. Gobin, There would appear to be a typo in the package manual that appears on CRAN. Doing ?Frechet (it would be nice to have this aliased also to frechet) points you immediately to frechet. With frechet2 substituted for frechet your code, everything works. cheers, Rolf Turner On 06/11/14 22:34, Katherine Gobin wrote: Dear Mr Michael, Thanks a lot for your guidance. The pdf file describing VGAM package has mentioned 'frechet' in the example, so I got the error. Regards Katherine On Thursday, 6 November 2014 2:54 PM, Michael Dewey i...@aghmed.fsnet.co.uk wrote: On 06/11/2014 06:04, Katherine Gobin wrote: Dear R forum, I am trying to execute following code (Page no 259 - VGAM.pdf) # . library(VGAM) set.seed(123) fdata - data.frame(y1 = rfrechet(nn - 1000, shape = 2 + exp(1))) with(fdata, hist(y1)) fit2 - vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) # . Is it not called frechet2? However, I receive following error Error in vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) : object 'frechet' not found Earlier there used to be a function called frechet3 which I guess has been withdrawn by VGAM. Kindly guide Katherine -- Rolf Turner Technical Editor ANZJS __ R-help@r-project.org 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] limit of cmdscale function
You avoid the call to cmdscale() by supplying your own starting configuration
(see the manual page for the y= argument). You could still hit other barriers
within isoMDS() or insufficient memory on your computer.
-
David L Carlson
Department of Anthropology
Texas AM University
College Station, TX 77840-4352
-Original Message-
From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On
Behalf Of Kawashima, Masayuki
Sent: Wednesday, November 5, 2014 10:51 PM
To: r-help@r-project.org
Subject: [R] limit of cmdscale function
Hi
We have a few questions regarding the use of the isoMDS function.
When we run isoMDS function using 60,000 x 60,000 data matrix,
we get the following error message:
cmdscale(d, k) : invalid value of 'n'
Calls: isoMDS - cmdscale
We checked the source code of cmdscale and found the following limitation:
## we need to handle nxn internally in dblcen
if(is.na(n) || n 46340) stop(invalid value of 'n')
1. This cmdscale limitation ('n 46340') is due to the limitation of BLAS and
LAPACK variables(int4) which can only handle '2^31-1' amount of data?
2. Is there any workaround to run isoMDS using large data (i.e. greater than
46340)?
We would like to run isoMDS using a maximum of 150,000x150,000 data matrix.
Best regards
Masayuki Kawashima
Email: kawasim...@jp.fujitsu.com
__
R-help@r-project.org 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@r-project.org 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] speed issue in simulating a stochastic process
I wish to simulate the following stochastic process, for i = 1...N
individuals and t=1...T periods:
y_{i,t} = y_0 + lambda Ey_{t-1} + epsilon_{i,t}
where Ey_{t-1} is the average of y over the N individuals computed at time
t-1.
My solution (below) works but is incredibly slow. Is there a faster but
still clear and readable alternative?
Thanks a lot. Matteo
rm(list=ls())
library(plyr)
y0 = 0
lambda = 0.1
N = 20
T = 100
m_e = 0
sd_e = 1
# construct the data frame and initialize y
D = data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
ybar - ddply(D,~t,summarise,mean=mean(y))
plot(ybar, col = blue, type = l)
[[alternative HTML version deleted]]
__
R-help@r-project.org 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] speed issue in simulating a stochastic process
Matteo,
I tried your example code using R 3.1.1 on an iMac (24-inch, Early 2009), 3.06
GHz Intel Core 2 Duo, 8 GB 1333 MHz DDR3, NVIDIA GeForce GT 130 512 MB
running Mac OS X 10.10 (Yosemite).
After entering your code, the elapsed time from the time I hit return to
when the graphics appeared was about 2 seconds — is this about what you are
seeing?
Regards,
Tom
On Thu, Nov 6, 2014 at 7:47 AM, Matteo Richiardi matteo.richia...@gmail.com
wrote:
I wish to simulate the following stochastic process, for i = 1...N
individuals and t=1...T periods:
y_{i,t} = y_0 + lambda Ey_{t-1} + epsilon_{i,t}
where Ey_{t-1} is the average of y over the N individuals computed at time
t-1.
My solution (below) works but is incredibly slow. Is there a faster but
still clear and readable alternative?
Thanks a lot. Matteo
rm(list=ls())
library(plyr)
y0 = 0
lambda = 0.1
N = 20
T = 100
m_e = 0
sd_e = 1
# construct the data frame and initialize y
D = data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
ybar - ddply(D,~t,summarise,mean=mean(y))
plot(ybar, col = blue, type = l)
[[alternative HTML version deleted]]
__
R-help@r-project.org 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.
[[alternative HTML version deleted]]
__
R-help@r-project.org 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] cluster + tt terms in coxph
On 2014-11-05 14:50, Therneau, Terry M., Ph.D. wrote: This is fixed in version 2.37-8 of the survival package, which has been in my send to CRAN real-soon-now queue for 6 months. Your note is a prod to get it done. I've been updating and adding vignettes. Is your fixed code publicly available somewhere? (The 'survival' repository at R-forge doesn't seem to have been updated since January.) Henric Winell Terry Therneau On 11/05/2014 05:00 AM, r-help-requ...@r-project.org wrote: I am receiving the following error when trying to include both tt (time transforms) and frailty terms in coxph coxph(Surv(time, status) ~ ph.ecog + tt(age)+cluster(sex), data=lung, + tt=function(x,t,...) pspline(x + t/365.25)) Error in residuals.coxph(fit2, type = dfbeta, collapse = cluster, weighted = TRUE) : Wrong length for 'collapse' I tried both 64 bit (R.3.1.0) and 32 bit (R.3.1.2) in Windows 7 64bit and get the same errors Inclusion of tt and cluster terms worked fine in R2.9.2-2.15.1 under Windows Vista 32 bit and Ubuntu 64 bit Any ideas? __ R-help@r-project.org 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@r-project.org 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] speed issue in simulating a stochastic process
Matteo,
Ah — OK, N=20, I did not catch that. You have nested for loops, which R is
known to be exceedingly slow at handling — if you can reorganize the code
to eliminate the loops, your performance will increase significantly.
Tom
On Thu, Nov 6, 2014 at 7:47 AM, Matteo Richiardi matteo.richia...@gmail.com
wrote:
I wish to simulate the following stochastic process, for i = 1...N
individuals and t=1...T periods:
y_{i,t} = y_0 + lambda Ey_{t-1} + epsilon_{i,t}
where Ey_{t-1} is the average of y over the N individuals computed at time
t-1.
My solution (below) works but is incredibly slow. Is there a faster but
still clear and readable alternative?
Thanks a lot. Matteo
rm(list=ls())
library(plyr)
y0 = 0
lambda = 0.1
N = 20
T = 100
m_e = 0
sd_e = 1
# construct the data frame and initialize y
D = data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
ybar - ddply(D,~t,summarise,mean=mean(y))
plot(ybar, col = blue, type = l)
[[alternative HTML version deleted]]
__
R-help@r-project.org 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.
[[alternative HTML version deleted]]
__
R-help@r-project.org 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] speed issue in simulating a stochastic process
I find that representing the simulated data as a T row by N column matrix
allows for a clearer and faster simulation function. E.g., compare the
output of the following two functions, the first of which uses your code
and the second a matrix representation (which I convert to a data.frame at
the end so I can compare outputs easily). I timed both of them for T=10^3
times and N=50 individuals; both gave the same results and f1 was 1
times faster than f0:
set.seed(1); t0 - system.time(s0 - f0(N=50,T=1000))
set.seed(1); t1 - system.time(s1 - f1(N=50,T=1000))
rbind(t0, t1)
user.self sys.self elapsed user.child sys.child
t0436.87 0.11 438.48 NANA
t1 0.04 0.000.04 NANA
all.equal(s0, s1)
[1] TRUE
The functions are:
f0 - function(N = 20, T = 100, lambda = 0.1, m_e = 0, sd_e = 1, y0 = 0)
{
# construct the data frame and initialize y
D - data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
D
}
f1 - function(N = 20, T = 100, lambda = 0.1, m_e = 0, sd_e = 1, y0 = 0)
{
# same process simulated using a matrix representation
# The T rows are times, the N columns are individuals
M - matrix(y0, nrow=T, ncol=N)
if (T 1) for(t in 2:T) {
ybar.L1 - mean(M[t-1L,])
epsilon - rnorm(N, mean=m_e, sd=sd_e)
M[t,] - lambda * y0 + (1-lambda)*ybar.L1 + epsilon
}
# convert to the data.frame representation that f0 uses
tM - t(M)
data.frame(id = as.vector(row(tM)), t = as.vector(col(tM)), y =
as.vector(tM))
}
Bill Dunlap
TIBCO Software
wdunlap tibco.com
On Thu, Nov 6, 2014 at 6:47 AM, Matteo Richiardi matteo.richia...@gmail.com
wrote:
I wish to simulate the following stochastic process, for i = 1...N
individuals and t=1...T periods:
y_{i,t} = y_0 + lambda Ey_{t-1} + epsilon_{i,t}
where Ey_{t-1} is the average of y over the N individuals computed at time
t-1.
My solution (below) works but is incredibly slow. Is there a faster but
still clear and readable alternative?
Thanks a lot. Matteo
rm(list=ls())
library(plyr)
y0 = 0
lambda = 0.1
N = 20
T = 100
m_e = 0
sd_e = 1
# construct the data frame and initialize y
D = data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
ybar - ddply(D,~t,summarise,mean=mean(y))
plot(ybar, col = blue, type = l)
[[alternative HTML version deleted]]
__
R-help@r-project.org 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.
[[alternative HTML version deleted]]
__
R-help@r-project.org 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] Daylight Saving Time
Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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] Timezone Upgrade
Hi, I would like to know, once the operating system timezone information is updated, what step should be carried out at R for reflecting the operating system(LINUX X64 SOLARIS X64 and WINDOWS X64) update. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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] Timezone Upgrade
Log out and log back in again. For many people this may happen in the normal course of daily use. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:28:37 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I would like to know, once the operating system timezone information is updated, what step should be carried out at R for reflecting the operating system(LINUX X64 SOLARIS X64 and WINDOWS X64) update. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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@r-project.org 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] Timezone Upgrade
I am log out and login many times. but no luck. Thanks Vasanth On Fri, Nov 7, 2014 at 1:04 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: Log out and log back in again. For many people this may happen in the normal course of daily use. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:28:37 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I would like to know, once the operating system timezone information is updated, what step should be carried out at R for reflecting the operating system(LINUX X64 SOLARIS X64 and WINDOWS X64) update. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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. [[alternative HTML version deleted]] __ R-help@r-project.org 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] Daylight Saving Time
R on Windows uses the Olsen timezone database, a copy of which is stored with R in the Program Files directory (e.g. R/R-3.1.1/share/zoneinfo). You could update the file yourself if you can find a corrected version, or download an updated version of R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:16:06 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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@r-project.org 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] Daylight Saving Time
Thanks Jeff Newmiller, it is working now. I would like to know, the same kind of configuration can be done in Linux and Solaris platform. Instead of R is mapping to operating system(/usr/share/. /usr/share/lib/) zoneinfo directory. Thanks Vasanth On Fri, Nov 7, 2014 at 1:18 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: R on Windows uses the Olsen timezone database, a copy of which is stored with R in the Program Files directory (e.g. R/R-3.1.1/share/zoneinfo). You could update the file yourself if you can find a corrected version, or download an updated version of R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:16:06 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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. [[alternative HTML version deleted]] __ R-help@r-project.org 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] Daylight Saving Time
Working now... after what action? AFAIK on *NIX systems R uses the OS installation of the Olsen database, so on a fresh login R should pick up any OS update you have installed. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 3:36:38 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Thanks Jeff Newmiller, it is working now. I would like to know, the same kind of configuration can be done in Linux and Solaris platform. Instead of R is mapping to operating system(/usr/share/. /usr/share/lib/) zoneinfo directory. Thanks Vasanth On Fri, Nov 7, 2014 at 1:18 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: R on Windows uses the Olsen timezone database, a copy of which is stored with R in the Program Files directory (e.g. R/R-3.1.1/share/zoneinfo). You could update the file yourself if you can find a corrected version, or download an updated version of R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:16:06 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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@r-project.org 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] Daylight Saving Time
In LINUX, I don't want R to use the operating system zoneinfo(Olsen database) instead of that I like to point different path which has the latest zoneinfo(latest Olsen database). On Fri, Nov 7, 2014 at 2:34 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: Working now... after what action? AFAIK on *NIX systems R uses the OS installation of the Olsen database, so on a fresh login R should pick up any OS update you have installed. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 3:36:38 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Thanks Jeff Newmiller, it is working now. I would like to know, the same kind of configuration can be done in Linux and Solaris platform. Instead of R is mapping to operating system(/usr/share/. /usr/share/lib/) zoneinfo directory. Thanks Vasanth On Fri, Nov 7, 2014 at 1:18 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: R on Windows uses the Olsen timezone database, a copy of which is stored with R in the Program Files directory (e.g. R/R-3.1.1/share/zoneinfo). You could update the file yourself if you can find a corrected version, or download an updated version of R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:16:06 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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. [[alternative HTML version deleted]] __ R-help@r-project.org 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] Daylight Saving Time
?timezones You probably need to recompile R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 4:10:58 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: In LINUX, I don't want R to use the operating system zoneinfo(Olsen database) instead of that I like to point different path which has the latest zoneinfo(latest Olsen database). On Fri, Nov 7, 2014 at 2:34 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: Working now... after what action? AFAIK on *NIX systems R uses the OS installation of the Olsen database, so on a fresh login R should pick up any OS update you have installed. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 3:36:38 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Thanks Jeff Newmiller, it is working now. I would like to know, the same kind of configuration can be done in Linux and Solaris platform. Instead of R is mapping to operating system(/usr/share/. /usr/share/lib/) zoneinfo directory. Thanks Vasanth On Fri, Nov 7, 2014 at 1:18 AM, Jeff Newmiller jdnew...@dcn.davis.ca.us wrote: R on Windows uses the Olsen timezone database, a copy of which is stored with R in the Program Files directory (e.g. R/R-3.1.1/share/zoneinfo). You could update the file yourself if you can find a corrected version, or download an updated version of R. --- Jeff NewmillerThe . . Go Live... DCN:jdnew...@dcn.davis.ca.usBasics: ##.#. ##.#. Live Go... Live: OO#.. Dead: OO#.. Playing Research Engineer (Solar/BatteriesO.O#. #.O#. with /Software/Embedded Controllers) .OO#. .OO#. rocks...1k --- Sent from my phone. Please excuse my brevity. On November 6, 2014 1:16:06 PM EST, Vasantha Kumar Kesavan info.vas...@gmail.com wrote: Hi, I am working R on windows 2012 R2 platform, I have updated the latest hotfixes for time zone information(Microsoft KB 2981580. http://support.microsoft.com/kb/2981580). But still R is not populating correct date time values(Standard and Daylight saving). Could you please advise me, how to enable R to pick up the latest time zone information/configurations. Example: R --vanilla Sys.setenv(TZ = America/Eirunepe); dt-c(seq(as.POSIXct(2013-11-09 20:00:00,tz=America/Eirunepe), as.POSIXct(2013-11-10 10:00:00 ,tz=America/Eirunepe), by=hour)); dt [1] 2013-11-09 20:00:00 AMT 2013-11-09 21:00:00 AMT [3] 2013-11-09 22:00:00 AMT 2013-11-09 23:00:00 AMT [5] 2013-11-10 00:00:00 AMT 2013-11-10 01:00:00 AMT [7] 2013-11-10 02:00:00 AMT 2013-11-10 03:00:00 AMT [9] 2013-11-10 04:00:00 AMT 2013-11-10 05:00:00 AMT [11] 2013-11-10 06:00:00 AMT 2013-11-10 07:00:00 AMT [13] 2013-11-10 08:00:00 AMT 2013-11-10 09:00:00 AMT [15] 2013-11-10 10:00:00 AMT For the “*America/Eirunepe*” time zone, the DST ended on Sun 10-Nov-2013 at 12:00:00 A.M. when local clocks were set backward 1 hour. as per the latest timezone configuration, the date time sequence should have 2013-11-09 23:00:00 twice. Thanks Vasanth [[alternative HTML version deleted]] __ R-help@r-project.org 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@r-project.org 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,
Re: [R] speed issue in simulating a stochastic process
Loops are not slow, but your code did a lot of unneeded operations in each
loop.
E.g, you computed
D$id==i D$t==t
for each row of D. That involves 2*nrow(D) equality tests for each of the
nrow(D)
rows, i.e., it is quadratic in N*T.
Then you did a data.frame replacement operation
D[k,]$y - newValue
where k is D$id==1D$t==t. This extracts the k'th row of D, then extracts
the 1-row 'y' column
from it, replaces it with the new value, then puts that row back into D.
If you must use
a data.frame, the equivalent
D$y[k] - newValue
is probably much faster (data.frames are lists of columns, so replacing a
column is fast).
Using a matrix to organize things is less flexible, but faster because you
don't have to search
when you want to find the element for a given id and time - you just do a
little arithmetic to
get the offset from the start of the matrix.
Bill Dunlap
TIBCO Software
wdunlap tibco.com
On Thu, Nov 6, 2014 at 2:05 PM, Matteo Richiardi matteo.richia...@gmail.com
wrote:
Hi William,
that's super. Thanks a lot. I knew that R is slow with loops, but did not
imagine so slow! B.t.w., what's the reason?
Final question: in your code you have mean(M[t-1L,]): what is the 'L'
for? I removed it at apparently the code produces the same output...
Thanks again,
Matteo
On 6 November 2014 18:46, William Dunlap wdun...@tibco.com wrote:
I find that representing the simulated data as a T row by N column matrix
allows for a clearer and faster simulation function. E.g., compare the
output of the following two functions, the first of which uses your code
and the second a matrix representation (which I convert to a data.frame at
the end so I can compare outputs easily). I timed both of them for T=10^3
times and N=50 individuals; both gave the same results and f1 was 1
times faster than f0:
set.seed(1); t0 - system.time(s0 - f0(N=50,T=1000))
set.seed(1); t1 - system.time(s1 - f1(N=50,T=1000))
rbind(t0, t1)
user.self sys.self elapsed user.child sys.child
t0436.87 0.11 438.48 NANA
t1 0.04 0.000.04 NANA
all.equal(s0, s1)
[1] TRUE
The functions are:
f0 - function(N = 20, T = 100, lambda = 0.1, m_e = 0, sd_e = 1, y0 = 0)
{
# construct the data frame and initialize y
D - data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
D
}
f1 - function(N = 20, T = 100, lambda = 0.1, m_e = 0, sd_e = 1, y0 = 0)
{
# same process simulated using a matrix representation
# The T rows are times, the N columns are individuals
M - matrix(y0, nrow=T, ncol=N)
if (T 1) for(t in 2:T) {
ybar.L1 - mean(M[t-1L,])
epsilon - rnorm(N, mean=m_e, sd=sd_e)
M[t,] - lambda * y0 + (1-lambda)*ybar.L1 + epsilon
}
# convert to the data.frame representation that f0 uses
tM - t(M)
data.frame(id = as.vector(row(tM)), t = as.vector(col(tM)), y =
as.vector(tM))
}
Bill Dunlap
TIBCO Software
wdunlap tibco.com
On Thu, Nov 6, 2014 at 6:47 AM, Matteo Richiardi
matteo.richia...@gmail.com wrote:
I wish to simulate the following stochastic process, for i = 1...N
individuals and t=1...T periods:
y_{i,t} = y_0 + lambda Ey_{t-1} + epsilon_{i,t}
where Ey_{t-1} is the average of y over the N individuals computed at
time
t-1.
My solution (below) works but is incredibly slow. Is there a faster but
still clear and readable alternative?
Thanks a lot. Matteo
rm(list=ls())
library(plyr)
y0 = 0
lambda = 0.1
N = 20
T = 100
m_e = 0
sd_e = 1
# construct the data frame and initialize y
D = data.frame(
id = rep(1:N,T),
t = rep(1:T, each = N),
y = rep(y0,N*T)
)
# update y
for(t in 2:T){
ybar.L1 = mean(D[D$t==t-1,y])
for(i in 1:N){
epsilon = rnorm(1,mean=m_e,sd=sd_e)
D[D$id==i D$t==t,]$y = lambda*y0+(1-lambda)*ybar.L1+epsilon
}
}
ybar - ddply(D,~t,summarise,mean=mean(y))
plot(ybar, col = blue, type = l)
[[alternative HTML version deleted]]
__
R-help@r-project.org 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.
[[alternative HTML version deleted]]
__
R-help@r-project.org 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] Differences between MTDFReml and kinship2
From: silvano silv...@uel.br I fitted a genetic model using kinship2 and I compared it with MTDFReml program output. The residual variance of both are very close but the genetic variance are very differents. The output are: MTDFReml: genetic variance = 1.24015 residual variance = 5.93424 R (Kinship2): genetic variance = 0.767187 residual variance = 5.6712 Both of them use REML method. Could someone tell why the difference? Hi Silvano. You haven't given us enough information, I'm afraid. We would need the commands you used, the versions of the software, plus output from both jobs, and your data, or a smaller subset that exhibits the same problem. You may have made an error, there may be a problem in one program, or the likelihood might be flat around those solutions (ie your data are not very informative). It might be better to continue this on the R-sig-mixed-models list. Cheers, David Duffy. __ R-help@r-project.org 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] VGAM package : Frechet distribution - 2 parameter estimation
Hello, a day or two ago I submitted VGAM 0.9-5 to CRAN, which has myriads of changes to family functions---their names, their arguments, and their order thereof. Especially regarding family functions for discrete and continuous distributions. In a nutshell, I found lots of inconsistencies while writing my book, and it was deemed necessary to standardize things. I just had to bite the bullet, so to speak. I have tried to summarize all changes in the NEWS file, however, a few might have gone undocumented. Users of previous versions of VGAM are cautioned to check their code. My apologies for this inconvenience. I have always tried to make it plain that while the version number was less than 1.0-0, everything was subject to change. But the book Vector Generalized Linear and Additive Models, for Springer, should appear next year, it is synchronized with version 1.0-0, so that should end most of these disruptive changes :) Or at least, curtail them (hopefully). cheers Thomas ps. for this particular problem, only frechet() remains. On 06/11/14 23:12, Rolf Turner wrote: Dear Ms. Gobin, There would appear to be a typo in the package manual that appears on CRAN. Doing ?Frechet (it would be nice to have this aliased also to frechet) points you immediately to frechet. With frechet2 substituted for frechet your code, everything works. cheers, Rolf Turner On 06/11/14 22:34, Katherine Gobin wrote: Dear Mr Michael, Thanks a lot for your guidance. The pdf file describing VGAM package has mentioned 'frechet' in the example, so I got the error. Regards Katherine On Thursday, 6 November 2014 2:54 PM, Michael Dewey i...@aghmed.fsnet.co.uk wrote: On 06/11/2014 06:04, Katherine Gobin wrote: Dear R forum, I am trying to execute following code (Page no 259 - VGAM.pdf) # . library(VGAM) set.seed(123) fdata - data.frame(y1 = rfrechet(nn - 1000, shape = 2 + exp(1))) with(fdata, hist(y1)) fit2 - vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) # . Is it not called frechet2? However, I receive following error Error in vglm(y1 ~ 1, frechet, data = fdata, trace = TRUE) : object 'frechet' not found Earlier there used to be a function called frechet3 which I guess has been withdrawn by VGAM. Kindly guide Katherine __ R-help@r-project.org 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] Lme4 Package Help!
Hello, all! So, as stated in the title, the Lme4 package used to output p-values for the fixed effects. What happened?! Literally 2 weeks ago, I ran code, got output with no errors, and had p-values listed for my fixed effects. Now, running THE SAME CODE with THE SAME DATASET (nothing at all has changed, not the data, not my computer, not R, nothing), I do not get p-values. I've tried other computers, I've tried resetting R. Any ideas? I'd really need to get some p-values. Thank you! D __ R-help@r-project.org 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] R 2.14.2 installation on mac OS X yosemite
Hi, I noticed your email on the R support page. I am currently using SPSS 21 for mac, and need to install a macro plug-in which requires me o install R 2.14 (I already have R 3.0). When i download the R 2.14 package from the CRAN website, and try to install it on my Macintosh HD, it gives me an error message : “ R 2.14.2 for Mac OS X 10.5 or higher (Leopard build) can’t be installed on this disk. Leopard build of R requires Mac OS X 10.5 or higher. Use Tiger build (available from CRAN) for older systems.” I dont understand this, since I have OS X Yosemite on my mac. I tried installing 2.14.0, but got the exact same error message. Would appreciate your help greatly. Deepansh __ R-help@r-project.org 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] R version 2.14.2 installation on mac OS X Yosemite
Hi, I noticed your email on the R support page. I am currently using SPSS 21 for mac, and need to install a macro plug-in which requires me o install R 2.14 (I already have R 3.0). When i download the R 2.14 package from the CRAN website, and try to install it on my Macintosh HD, it gives me an error message : “ R 2.14.2 for Mac OS X 10.5 or higher (Leopard build) can’t be installed on this disk. Leopard build of R requires Mac OS X 10.5 or higher. Use Tiger build (available from CRAN) for older systems.” I dont understand this, since I have OS X Yosemite on my mac. I tried installing 2.14.0, but got the exact same error message. Would appreciate your help greatly. Deepansh CONFIDENTIALITY NOTICE: This email contains information from the sender that may be CONFIDENTIAL, LEGALLY PRIVILEGED, PROPRIETARY or otherwise protected from disclosure. This email is intended for use only by the person or entity to whom it is addressed. If you are not the intended recipient, any use, disclosure, copying, distribution, printing, or any action taken in reliance on the contents of this email, is strictly prohibited. If you received this email in error, please contact the sending party by reply email, delete the email from your computer system and shred any paper copies. Note to Patients: There are a number of risks you should consider before using e-mail to communicate with us. See our Privacy Security page on www.henryford.com for more detailed information as well as information concerning MyChart, our new patient portal. If you do not believe that our policy gives you the privacy and security protection you need, do not send e-mail or Internet communications to us. __ R-help@r-project.org 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] R 2.14.2 installation on mac OS X yosemite
On 06.11.2014 20:21, Deepansh Dalela wrote: Hi, I noticed your email on the R support page. I am currently using SPSS 21 for mac, and need to install a macro plug-in which requires me o install R 2.14 (I already have R 3.0). When i download the R 2.14 package from the CRAN website, and try to install it on my Macintosh HD, it gives me an error message : “ R 2.14.2 for Mac OS X 10.5 or higher (Leopard build) can’t be installed on this disk. Leopard build of R requires Mac OS X 10.5 or higher. Use Tiger build (available from CRAN) for older systems.” I dont understand this, since I have OS X Yosemite on my mac. I tried installing 2.14.0, but got the exact same error message. Such old versions of R are not supported, particularly not on more recent OS. If you need help, ask the company for support that tells you to install an obsolete version of R. Best, Uwe Ligges Would appreciate your help greatly. Deepansh __ R-help@r-project.org 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@r-project.org 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] Lme4 Package Help!
Daniel Mello dmello2 at ucmerced.edu writes: Hello, all! So, as stated in the title, the Lme4 package used to output p-values for the fixed effects. What happened?! Literally 2 weeks ago, I ran code, got output with no errors, and had p-values listed for my fixed effects. Now, running THE SAME CODE with THE SAME DATASET (nothing at all has changed, not the data, not my computer, not R, nothing), I do not get p-values. I've tried other computers, I've tried resetting R. Any ideas? I'd really need to get some p-values. Thank you! D Applying my mind-reading skills, I'm going to guess that you are running anova() and that you had previously loaded the lmerTest package (which extends the anova() method from lme4 in several ways), and that now you are loading only the base lme4 package and not the lmerTest package. If my guess isn't right, then you're going to have to provide more information. Follow-ups to r-sig-mixed-mod...@r-project.org , please . Ben Bolker __ R-help@r-project.org 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] R version 2.14.2 installation on mac OS X Yosemite
On 06.11.2014 20:20, Dalela, Deepansh wrote: Hi, I noticed your email on the R support page. I am currently using SPSS 21 for mac, and need to install a macro plug-in which requires me o install R 2.14 (I already have R 3.0). When i download the R 2.14 package from the CRAN website, and try to install it on my Macintosh HD, it gives me an error message : “ R 2.14.2 for Mac OS X 10.5 or higher (Leopard build) can’t be installed on this disk. Leopard build of R requires Mac OS X 10.5 or higher. Use Tiger build (available from CRAN) for older systems.” I dont understand this, since I have OS X Yosemite on my mac. I tried installing 2.14.0, but got the exact same error message. Would appreciate your help greatly. Such old versions of R are not supported, particularly not on more recent OS. If you need help, ask the company for support that tells you to install an obsolete version of R. Best, Uwe Ligges Deepansh CONFIDENTIALITY NOTICE: This email contains information from the sender that may be CONFIDENTIAL, LEGALLY PRIVILEGED, PROPRIETARY or otherwise protected from disclosure. This email is intended for use only by the person or entity to whom it is addressed. If you are not the intended recipient, any use, disclosure, copying, distribution, printing, or any action taken in reliance on the contents of this email, is strictly prohibited. If you received this email in error, please contact the sending party by reply email, delete the email from your computer system and shred any paper copies. Note to Patients: There are a number of risks you should consider before using e-mail to communicate with us. See our Privacy Security page on www.henryford.com for more detailed information as well as information concerning MyChart, our new patient portal. If you do not believe that our policy gives you the privacy and security protection you need, do not send e-mail or Internet communications to us. __ R-help@r-project.org 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@r-project.org 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] speed issue in simulating a stochastic process
SNIP On Thu, Nov 6, 2014 at 2:05 PM, Matteo Richiardi matteo.richia...@gmail.com wrote: SNIP Final question: in your code you have mean(M[t-1L,]): what is the 'L' for? I removed it at apparently the code produces the same output... SNIP The constant 1L is stored as an integer; the constant 1 is stored as double precision. This sometimes makes no difference and sometimes makes a huge difference (especially in the context of numerical comparisons). If something is supposed to be an integer it is safer to use the L form. See ?NumericConstants. cheers, Rolf Turner -- Rolf Turner Technical Editor ANZJS __ R-help@r-project.org 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-es] Solo enteros en el eje Y de graficos con lattice
Hola, Los ejes los controlas con scales. Es una lista, y uno de sus componentes puede ser y. Es a su vez otra lista en la que puedes configurar el eje con detalle. En help(xyplot) está descrito en un apartado que empieza así: scales: Generally a list determining how the x- and y-axes (tick marks and labels) are drawn. Por ejemplo: xyplot(1:3 ~1:3, scales = list(y = list(at = 1:3))) Saludos. Oscar. - Oscar Perpiñán Lamigueiro Dpto. Ingeniería Eléctrica (ETSIDI-UPM) Grupo de Sistemas Fotovoltaicos (IES-UPM) URL: http://oscarperpinan.github.io ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] VI Jornadas de Usuarios de R [Materiales]
Me uno a los agradecidos! El 6 de noviembre de 2014, 10:35, Lorena Tudela Marco lorenatudelama...@gmail.com escribió: Muchas gracias por compartir! ;) Un abrazo, y buen día. Lore El 4 de noviembre de 2014, 22:33, Igor Sosa Mayor joseleopoldo1...@gmail.com escribió: miguel.angel.rodriguez.mui...@sergas.es writes: Hola a todos. Ya están disponibles los materiales (videos y presentaciones) de las VI Jornadas de Usuarios de R que se celebraron en Santiago de Compostela los días 23 y 24 de Octubre. http://r-es.org/Programa+de+las+VI+Jornadas mil gracias! -- :: Igor Sosa Mayor :: joseleopoldo1...@gmail.com :: :: GnuPG: 0x1C1E2890 :: http://www.gnupg.org/ :: :: jabberid: rogorido :::: ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] Duda_Observed vs Predicted
Estimada Lorena Tudela Voy a escribir algo por como yo lo aprendí (quizás mal redactado para un estadśitico), lo observado es lo que se mide a campo y lo predicho es lo que surge a partir del modelo estadístico. Luego aprendí como realizarlo con R, pero no siempre es posible en forma fácil, porque hay librerías que tienen dentro de sus algoritmos la forma adecuada para la predicción, en cambio otras no, y hay que realizarlo a mano. Por lo cuál es necesario conocer que herramienta dentro de R está utilizando para sus modelos. Porque, puede ser que ella misma tenga la solución o se deba escribir en código R, y no siempre se pueden combinar una librería con la otra. Javier Marcuzzi El 06/11/14 a las 06:48, Lorena Tudela Marco escibió: Buenos días a todxs, Estoy comparando la predicción de los valores (0, 1, 2, 3,.hasta 13) frente a los observados. Con la idea de comparar el modelo Zero inflated y el Binomial negativo y ver cual presenta mas distancia frente a las predicciones observadas. Para ello introduzco los códigos en la consola: #Modelo ZIM pred-round(colSums(predict(zeroinfl, type=prob) [,1:14])) #Valores observados realmente obs-table(IB$nijt)[1:14] #Tabla comparativa rbind( pred, obs) Y obtengo la siguiente tabla: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 pred 3600 589 349 224 151 105 75 55 42 32 25 20 16 13 obs 3529 743 300 203 135 81 76 44 33 37 30 12 14 13 La duda me surge al intentarlo con el Modelo Binomial negativo.¿Sabeis que comando podría introducir para obtener los 13 valores predichos por el modelo BN? Muchas gracias por vuestra ayuda y buen día. Lorena [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] Solo enteros en el eje Y de graficos con lattice
Funciono perfecto, muchas gracias Oscar. Debo decir, que lei esa parte de la ayuda de xyplot, pero jamas se me habria ocurrido que servia para controlar lo que necesitaba. Slds y gracias, Eric. On Thu 06 Nov 2014 05:49:06 CLST, Oscar Perpiñan wrote: Hola, Los ejes los controlas con scales. Es una lista, y uno de sus componentes puede ser y. Es a su vez otra lista en la que puedes configurar el eje con detalle. En help(xyplot) está descrito en un apartado que empieza así: scales: Generally a list determining how the x- and y-axes (tick marks and labels) are drawn. Por ejemplo: xyplot(1:3 ~1:3, scales = list(y = list(at = 1:3))) Saludos. Oscar. - Oscar Perpiñán Lamigueiro Dpto. Ingeniería Eléctrica (ETSIDI-UPM) Grupo de Sistemas Fotovoltaicos (IES-UPM) URL: http://oscarperpinan.github.io -- Forest Engineer Master in Environmental and Natural Resource Economics Ph.D. student in Sciences of Natural Resources at La Frontera University Member in AguaDeTemu2030, citizen movement for Temuco with green city standards for living Nota: Las tildes se han omitido para asegurar compatibilidad con algunos lectores de correo. ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] Duda_Observed vs Predicted
Lorena Ni idea: Parte 1, nlme puede calcular con predict..., tendría que leer las especificaciones de nlme, porque dentro de las opciones puede ser que ... Parte 2, tantas librerías juntas, ..., puede ser que la última que carga sobrepone una función a las anteriores. Yo no puedo opinar sobre librerías que nunca use, de pronto lo que usted escribe es correcto, pero yo no puedo asegurarlo, el problema me supera. Javier El 06/11/14 a las 12:42, Lorena Tudela Marco escibió: Hola Javier, Si, cuando hablo de valor observado me refiero al valor real en campo y el predicho al que estiman los modelos. Disculpa, que no lo detallase así desde el principio. En mi caso trabajo con dos diferentes: Zero inflated y Binomial Negativo y me gustaría comprobar que diferencia (distancia) existe entre cada uno de ellos y la realidad. Estoy trabajando con los siguientes paquetes: library(pscl) library(MASS) library(AER) library(VGAM) library(truncreg) library(censReg) library(sampleSelection) library(ggplot2) library(boot) library(aod) library(lmtest) library(zoo) library(nlme) library(lmtest) library(boot) library(spatcounts) Mi duda es: ¿Que comando podría utilizar para calcular por valores predichos por el Binomial negativo para los 14 primeros valores de la variable dependiente? En el caso del ZIM utilizo: pred-round(colSums(predict(zeroinfl, type=prob) [,1:14])) Gracias por vuestra ayuda! ;) Lore El 6 de noviembre de 2014, 12:30, Marcuzzi, Javier Rubén javier.ruben.marcu...@gmail.com mailto:javier.ruben.marcu...@gmail.com escribió: Estimada Lorena Tudela Voy a escribir algo por como yo lo aprendí (quizás mal redactado para un estadśitico), lo observado es lo que se mide a campo y lo predicho es lo que surge a partir del modelo estadístico. Luego aprendí como realizarlo con R, pero no siempre es posible en forma fácil, porque hay librerías que tienen dentro de sus algoritmos la forma adecuada para la predicción, en cambio otras no, y hay que realizarlo a mano. Por lo cuál es necesario conocer que herramienta dentro de R está utilizando para sus modelos. Porque, puede ser que ella misma tenga la solución o se deba escribir en código R, y no siempre se pueden combinar una librería con la otra. Javier Marcuzzi El 06/11/14 a las 06:48, Lorena Tudela Marco escibió: Buenos días a todxs, Estoy comparando la predicción de los valores (0, 1, 2, 3,.hasta 13) frente a los observados. Con la idea de comparar el modelo Zero inflated y el Binomial negativo y ver cual presenta mas distancia frente a las predicciones observadas. Para ello introduzco los códigos en la consola: #Modelo ZIM pred-round(colSums(predict(zeroinfl, type=prob) [,1:14])) #Valores observados realmente obs-table(IB$nijt)[1:14] #Tabla comparativa rbind( pred, obs) Y obtengo la siguiente tabla: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 pred 3600 589 349 224 151 105 75 55 42 32 25 20 16 13 obs 3529 743 300 203 135 81 76 44 33 37 30 12 14 13 La duda me surge al intentarlo con el Modelo Binomial negativo.¿Sabeis que comando podría introducir para obtener los 13 valores predichos por el modelo BN? Muchas gracias por vuestra ayuda y buen día. Lorena [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] Duda_Observed vs Predicted
Hola Lorena. Trabajé hace tiempo con binomiales negativas y esta página me sirvió bastante: http://www.ats.ucla.edu/stat/r/dae/nbreg.htm Ahí tienes una manera de sacar los predicted values para una negbinom. No se si es exactamente lo que buscas, pero te puede servir como punto de partida para encontrar la solución en caso de que no lo sea. Espero que te sirva. Un saludo. Víctor Granda García Ph.D. Student Dpto. BOS, Universidad de Oviedo El 6 de noviembre de 2014, 18:06, Marcuzzi, Javier Rubén javier.ruben.marcu...@gmail.com escribió: Lorena Ni idea: Parte 1, nlme puede calcular con predict..., tendría que leer las especificaciones de nlme, porque dentro de las opciones puede ser que ... Parte 2, tantas librerías juntas, ..., puede ser que la última que carga sobrepone una función a las anteriores. Yo no puedo opinar sobre librerías que nunca use, de pronto lo que usted escribe es correcto, pero yo no puedo asegurarlo, el problema me supera. Javier El 06/11/14 a las 12:42, Lorena Tudela Marco escibió: Hola Javier, Si, cuando hablo de valor observado me refiero al valor real en campo y el predicho al que estiman los modelos. Disculpa, que no lo detallase así desde el principio. En mi caso trabajo con dos diferentes: Zero inflated y Binomial Negativo y me gustaría comprobar que diferencia (distancia) existe entre cada uno de ellos y la realidad. Estoy trabajando con los siguientes paquetes: library(pscl) library(MASS) library(AER) library(VGAM) library(truncreg) library(censReg) library(sampleSelection) library(ggplot2) library(boot) library(aod) library(lmtest) library(zoo) library(nlme) library(lmtest) library(boot) library(spatcounts) Mi duda es: ¿Que comando podría utilizar para calcular por valores predichos por el Binomial negativo para los 14 primeros valores de la variable dependiente? En el caso del ZIM utilizo: pred-round(colSums(predict(zeroinfl, type=prob) [,1:14])) Gracias por vuestra ayuda! ;) Lore El 6 de noviembre de 2014, 12:30, Marcuzzi, Javier Rubén javier.ruben.marcu...@gmail.com mailto:javier.ruben.marcu...@gmail.com escribió: Estimada Lorena Tudela Voy a escribir algo por como yo lo aprendí (quizás mal redactado para un estadśitico), lo observado es lo que se mide a campo y lo predicho es lo que surge a partir del modelo estadístico. Luego aprendí como realizarlo con R, pero no siempre es posible en forma fácil, porque hay librerías que tienen dentro de sus algoritmos la forma adecuada para la predicción, en cambio otras no, y hay que realizarlo a mano. Por lo cuál es necesario conocer que herramienta dentro de R está utilizando para sus modelos. Porque, puede ser que ella misma tenga la solución o se deba escribir en código R, y no siempre se pueden combinar una librería con la otra. Javier Marcuzzi El 06/11/14 a las 06:48, Lorena Tudela Marco escibió: Buenos días a todxs, Estoy comparando la predicción de los valores (0, 1, 2, 3,.hasta 13) frente a los observados. Con la idea de comparar el modelo Zero inflated y el Binomial negativo y ver cual presenta mas distancia frente a las predicciones observadas. Para ello introduzco los códigos en la consola: #Modelo ZIM pred-round(colSums(predict(zeroinfl, type=prob) [,1:14])) #Valores observados realmente obs-table(IB$nijt)[1:14] #Tabla comparativa rbind( pred, obs) Y obtengo la siguiente tabla: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 pred 3600 589 349 224 151 105 75 55 42 32 25 20 16 13 obs 3529 743 300 203 135 81 76 44 33 37 30 12 14 13 La duda me surge al intentarlo con el Modelo Binomial negativo.¿Sabeis que comando podría introducir para obtener los 13 valores predichos por el modelo BN? Muchas gracias por vuestra ayuda y buen día. Lorena [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es [[alternative HTML version deleted]] ___ R-help-es mailing list
Re: [R-es] diferentes escalas en el X de un grafico con varios paneles
Hola Gerardo, antes hice algo parecido para una presentacion a publico general, y nadie se dio cuenta, pero ahora tengo que enviar ese grafico a una revista cientifica. Gracias. Slds, eric. On Thu 06 Nov 2014 14:20:39 CLST, Gerardo Gold-Bouchot wrote: Puedes sumarle 1 a todos tus valores, para que puedas usar la escala logarítmica. Saludos, Gerardo 2014-11-06 11:07 GMT-06:00 eric ericconchamu...@gmail.com mailto:ericconchamu...@gmail.com: Estimados, tengo el siguiente problema: Tengo que graficar la concentracion de un metabolito en algunas muestras contra la concentracion del solvente con que fue extraido. Las concentraciones varian desde cero (para un control) hasta varios miles de unidades de concentracion (miliMolar). Para que se puedan diferenciar los puntos en el eje X, se ha usado una escala logaritmica. Hasta ahi todo bien. El problema se produce con los valores del control, que por tener una concentracion de solvente igual a cero, no puede aparecer en el grafico, pues, como saben, la escala logaritmica no admite un cero. Como podria hacer un grafico como el que se adjunta pero incluyendo los datos del control y manteniendo la escala logaritmica en el eje X ? Alguna idea ? Haaa aprovecho de preguntar porque aparecen dos marcas tan juntas en el grafico que adjunto. Este es el codigo con el que lo hice: xyplot(V1 ~ con | sol + dia , groups=nca , data=de.pgc , auto.key=list(space=right) , strip = strip.custom(strip.names = TRUE) , scales=list(x=list(log=10), equispaced.log=FALSE, y = list(at = 1:3), cex=1.2) , xlab=list(Solvent concentration (mM), cex=1.2) , ylab=list(Numbers of samples with cis bonds, cex=1.2) , ylim=c(0.5,3.5) #, xlim=c(0,5000) ) Slds, eric. -- Forest Engineer Master in Environmental and Natural Resource Economics Ph.D. student in Sciences of Natural Resources at La Frontera University Member in AguaDeTemu2030, citizen movement for Temuco with green city standards for living Nota: Las tildes se han omitido para asegurar compatibilidad con algunos lectores de correo. ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es -- Forest Engineer Master in Environmental and Natural Resource Economics Ph.D. student in Sciences of Natural Resources at La Frontera University Member in AguaDeTemu2030, citizen movement for Temuco with green city standards for living Nota: Las tildes se han omitido para asegurar compatibilidad con algunos lectores de correo. ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
Re: [R-es] diferentes escalas en el X de un grafico con varios paneles
Es un truco muy com�n en varias disciplinas, particularmente en Ecolog�a, as� que no deber�a haber problema. De hecho en la Biblia de Ecolog�a Cuantitativa, de Legendre y Legendre, se recomienda sumar un n�mero que sea del mismo orden de magnitud del valor de la media op mediana de la variable, para no sesgar los resultados hacia arriba o hacia abajo, pero dado que tu solo quieres un gr�fico creo que sumar 1 es ideal. Saludos, Gerardo 2014-11-06 11:48 GMT-06:00 eric ericconchamu...@gmail.com: Hola Gerardo, antes hice algo parecido para una presentacion a publico general, y nadie se dio cuenta, pero ahora tengo que enviar ese grafico a una revista cientifica. Gracias. Slds, eric. On Thu 06 Nov 2014 14:20:39 CLST, Gerardo Gold-Bouchot wrote: Puedes sumarle 1 a todos tus valores, para que puedas usar la escala logar�tmica. Saludos, Gerardo 2014-11-06 11:07 GMT-06:00 eric ericconchamu...@gmail.com mailto:ericconchamu...@gmail.com: Estimados, tengo el siguiente problema: Tengo que graficar la concentracion de un metabolito en algunas muestras contra la concentracion del solvente con que fue extraido. Las concentraciones varian desde cero (para un control) hasta varios miles de unidades de concentracion (miliMolar). Para que se puedan diferenciar los puntos en el eje X, se ha usado una escala logaritmica. Hasta ahi todo bien. El problema se produce con los valores del control, que por tener una concentracion de solvente igual a cero, no puede aparecer en el grafico, pues, como saben, la escala logaritmica no admite un cero. Como podria hacer un grafico como el que se adjunta pero incluyendo los datos del control y manteniendo la escala logaritmica en el eje X ? Alguna idea ? Haaa aprovecho de preguntar porque aparecen dos marcas tan juntas en el grafico que adjunto. Este es el codigo con el que lo hice: xyplot(V1 ~ con | sol + dia , groups=nca , data=de.pgc , auto.key=list(space=right) , strip = strip.custom(strip.names = TRUE) , scales=list(x=list(log=10), equispaced.log=FALSE, y = list(at = 1:3), cex=1.2) , xlab=list(Solvent concentration (mM), cex=1.2) , ylab=list(Numbers of samples with cis bonds, cex=1.2) , ylim=c(0.5,3.5) #, xlim=c(0,5000) ) Slds, eric. -- Forest Engineer Master in Environmental and Natural Resource Economics Ph.D. student in Sciences of Natural Resources at La Frontera University Member in AguaDeTemu2030, citizen movement for Temuco with green city standards for living Nota: Las tildes se han omitido para asegurar compatibilidad con algunos lectores de correo. ___ R-help-es mailing list R-help-es@r-project.org mailto:R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es -- Forest Engineer Master in Environmental and Natural Resource Economics Ph.D. student in Sciences of Natural Resources at La Frontera University Member in AguaDeTemu2030, citizen movement for Temuco with green city standards for living Nota: Las tildes se han omitido para asegurar compatibilidad con algunos lectores de correo. [[alternative HTML version deleted]] ___ R-help-es mailing list R-help-es@r-project.org https://stat.ethz.ch/mailman/listinfo/r-help-es
