Hi Todd,
I have played around with a few specifications, but I can't seem to
replicate this error on my machine. I take it that when you do not set
the bounds, Amelia runs fine? If this is true, perhaps you could lower
the number of variables on which you set bounds.
Another tip is that these bounds are largely unnecessary unless there
is a logical constraint in your data and non-bounded runs of Amelia do
not meet these bounds. For instance, they are not intended to be used
when imputations seem to be "too varied" since this variation is part
of valid inference from imputed data.
If none of the above applies to you and you feel comfortable, you can
email me (off-list) with your data and I can try to pin down the exact
location of the problem. You would, of course, have a guarantee of
non-disclosure of your data (outside of the Amelia team) along with a
promise to remove the data once the problem is resolved.
I hope this helps,
matt.
On Mon, Apr 27, 2009 at 3:26 PM, Todd Whitehead
<toddpwhitehead(a)gmail.com> wrote:
Hi Matt,
I tried to update to the newest version of Amelia, but that did not seem to
solve my problem. Please let me know if I can help you figure out what is
going wrong.
Thanks,
Todd
On Mon, Apr 27, 2009 at 11:55 AM, Matt Blackwell <mblackwell(a)gmail.com>
wrote:
Hi Todd,
I think there may have been a bug in the "bounds" argument that caused
this problem. I've updated Amelia and it should be fixed. Since it's
not yet on CRAN, you can update to this newest version by running the
following line of code:
install.packages("Amelia", repos="http://gking.harvard.edu")
If you still have problems with the bounds, let us know and we can
investigate further.
Thanks for using Amelia,
matt.
On Mon, Apr 27, 2009 at 2:01 PM, Todd Whitehead
<toddpwhitehead(a)gmail.com> wrote:
> Hi,
>
> I'm using Amelia in R to impute missing values for my data, but I'm
> having a
> problem. When I include a bounds= argument, I get an error message as
> follows:
>
> "Error in matrix(NA, nrow = AMn.ss, ncol = AMp) :
> non-numeric matrix extent"
>
> Here is what my code looks like:
>
> imput.bound<-matrix(c(1,2,3,4,5,6,7,8,9,10,
> -50,-50,-50,-50,-50,-50,-50,-50,-50,-50,
> 20,20,20,20,20,20,20,20,20,20),
> nrow = 10, ncol=3, byrow=FALSE,)
> blod<-amelia(data=nonmiss, m=5, p2s=1, idvars="childid",
> bounds=imput.bound,
> max.resample=100)
>
> Does anybody see what I'm doing wrong? I tried looking through the
> reference guide, "Amelia II: A Program for Missing Data", but I
couldn't
> find any info on this particular error message. Any advice you could
> offer
> would really be appreciated.
>
> Thanks,
> Todd
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: