Hi Felipe, 

The problem with the data you are using is that there are rows that are completely missing. That is, there is no observed information for those rows. Amelia cannot impute rows with no observed data, so it returns those rows with all NAs. There is a warning message about this when you run Amelia, but you may have missed it at the beginning of the imputation procedure. 

Cheers,
matt. 

On Tue, Aug 2, 2011 at 12:32 PM, Luis Felipe Parra <felipe.parra@quantil.com.co> wrote:
Hello Matt, thanks for your answer, the code I am using is the following

path = "C:\\Users\\Hp\\Desktop"
load(paste(path,"\\y.rdata",sep=""))

out = amelia(y)#[4000:5000,])
summary(y)#[4000:5000,])
for(i in 1:length(out$imputations)){ print(summary(out$imputations[[i]])) }

and the data I send it attached. Thanks again for any help you can give me with this

Felipe Parra


On Tue, Aug 2, 2011 at 11:23 AM, Matt Blackwell <blackwel@fas.harvard.edu> wrote:
Hi Felipe, 

Without more information about the data and the exact code you are using to impute it, it is hard to know what is happening. It could be the case that the NAs in your data are not coded correctly and therefore not recognized as missing by R. You can check this by using:

sum(is.na(a.out$imputations[[1]]))

This code will tell you how many missing observations there are in the (supposedly) imputed dataset. This number might be greater than zero if you tagged certain variables as "idvars"---these variables are not imputed. 

Cheers,
matt. 


On Tue, Aug 2, 2011 at 11:48 AM, Luis Felipe Parra <felipe.parra@quantil.com.co> wrote:
Hello, Thank you Matt. I am doing exactly this 
a.out$imputations[[1]]
but I'am still getting NA's on this imputations. Do you know which migh be the reason?

Thank you

Felipe Parra


On Tue, Aug 2, 2011 at 10:19 AM, Matt Blackwell <blackwel@fas.harvard.edu> wrote:
Hi Luis, 

You want to make sure that you save the output to a name:

a.out <- amelia(y)

Once you do this, you can see your imputed datasets here:

a.out$imputations[[1]]

That is, for example, the first imputed dataset. You can access others by changing the "1" to another number. Hope that helps. 

Cheers,
matt.


On Thu, Jul 28, 2011 at 1:03 PM, Luis Felipe Parra <felipe.parra@quantil.com.co> wrote:
Hello, I've just discovered amelia today and I am trying to use to to fill some missing values in the attached dataset. I am using amelia in R, when I call amelia(y) I am getting my output with still some missing values.

Does somebody know what might be I doing wrong, or what might be going on. Thank you

Felipe Parra