Hi,
I am trying to produce tscsPlots to compare the output of different Amelia models and every time I run the tscsPlot command on the same output I get a different graph (i.e. different values on the graph). Is this how it is supposed to be? In the Amelia vignette it says that the red points are the mean imputation for each of the missing values, but is this the mean of the 5 originally imputed values (if m=5), or is it the mean of the 100 imputations that tscsPlot runs?
Any help will be much appreciated!
Thanks,
Caterina
Is there any way to get a progress bar for the imputation progress in the
command line version of Amelia? I'd like to be able to estimate how long
the whole imputation process will take. For loops, one can use the
following for a progress bar:
pb <- txtProgressBar(min = 0, max = 999999, style = 3)
for (i in 1:999999){
setTxtProgressBar(pb, i)
}
Is there an option to get a progress bar for a single or multiple
imputations in the command line version of Amelia?
Thanks!
Hi folks,
James, Gary, and I are happy to announce Amelia 1.7 is up on CRAN for your
enjoyment. You can install the update from the R prompt with the following
line of code:
install.packages("Amelia", repos = "http://cran.r-project.org")
We have also updated the Windows installer at:
http://gking.harvard.edu/amelia
Amelia 1.7 brings a fairly large update to Amelia, which includes a
complete rewrite of the Amelia internals to Rcpp. What does this mean for
you? Much, much faster imputations (we have seen speedups of 10x on some
examples). We have also added support for running Amelia in parallel using
multiple cores in modern CPUs. You can find more about parallel in section
4.3 on page 16 of the Amelia manual:
http://cran.r-project.org/web/packages/Amelia/vignettes/amelia.pdf
Overall, this version should make imputations much faster.
Other improvements and bugfixes:
* Plots in AmeliaView should now use Quartz on Mac OS X instead of X11.
* Amelia now requires R >=2.14.0.
* Amelia now can run its imputations in parallel using infrastructure
from R's parallel package. Note that R < 2.15.3 will crash if parallel is
used while tcltk is loaded (or has been loaded and then unloaded). This
will be fixed in R 2.15.3 (the patched version of 2.15.2) and we will
require R>=2.15.3 when that version is released.
* Fixed bug with character variables set to nominal.
Please let us know if you run into any problems with this new version.
Cheers,
matt.
~~~~~~~~~~~
Matthew Blackwell
Assistant Professor of Political Science
University of Rochester
url: http://www.mattblackwell.org
Dear list,
I installed 1.7 from source from the Cran repo to try out parallel
imputation.
Before 1.7 I did it with foreach
mtfs_i <- foreach(i=1:5,.combine="ameliabind",.inorder=F) %dopar% {
amelia(mtfs,
m = 1,
)
}
and it worked fine.
with 1.7 I tried using multicore using the parallel argument.
mtfs_i <- amelia(mtfs,
m = 1,
parallel = "multicore", # note: this will crash R if tcltk is loaded prior to 2.15.3
ncpus = parallel::detectCores(),
)
Both approaches finish without error, but I get a messed-up S3 object.
So, in approach 1 using foreach ameliabind complains that it didn't get amelia objects.
With approach 2 I get a list of Amelia objects, but the first two items are NULL and I don't get one list $imputations with many imputations, but a stitched-up object.
Using this object with e.g. transform or overimpute fails, but doing the same with the third object in the list works (that's just one imputation then though).
Regards,
Ruben Arslan
student assistant
Lab: http://www.psychology.hu-berlin.de/profship/perdev
Humboldt-University of Berlin
Unter den Linden 6
10099 Berlin, Germany
I would like to impute a variable that counts the number of delinquent acts
an individual engaged in in the last year. The variable will be used as an
outcome and has a 0-inflated distribution so it will need to have all
integer values if it is to be analyzed in a 0 inflated model.
Does anyone have thoughts on whether it's better to 1. impute this variable
as ordinal in order to obtain all integer values versus 2. impute as
continuous and recode the values to integers after imputation?
--
Alicia Doyle Lynch, Ph.D.
Boston College, Lynch School of Education
Department of Developmental and Educational Psychology
Chestnut Hill, MA 02467
Phone: (617) 552-6437
e-mail: doylead(a)bc.edu
Dear Lister,
I am using Amelia II (Version 1.6.4) with a 500 GB computer specification
and my data consist of directed dyads and my imputation model has 94
variables and 493,853 observations. I use the following command:
library(Amelia)
library(foreign)
mydata <- read.dta("data.dta")
require(Amelia)
set.seed(1234)
a.out <- amelia(mydata, m=10, p2s = 2, tolerance = 0.005, empri =
.1*nrow(mydata), ts="year", cs="dyadid" , polytime=2, intercs = TRUE)
After 7 hours, I receive the following message:
amelia starting
beginning prep functions
*Error in cbind(deparse.level, ...) :*
* resulting vector exceeds vector length limit in 'AnswerType'*
I've already searched the Amelia II archieves and R archives, I was not
able to locate a solution.
I would deeply appreciate any help!
Best Regards,
Ömer
_________________________________
Ömer Faruk Örsün
PhD Candidate
Department of International Relations
Koç University
CAS 289
_________________________________