Hello,
I have a dataset with some missing values, I used Amelia on this dataset
and ended up with 5 imputed datasets. I calculated the correlation matrix
for each dataset, therefore I have 5 correlation matrices. I would like to
know how to combine these 5 matrices in order to get one final correlation
matrix for my dataset?
Thank you very much
ERV
Dear Amelia Initiator,
This is Yunyan Zhang. I am a senior student majoring in Industrial and
Operations Engineering at the University of Michigan. Our team is doing our
senior design and we want to use R to deal with our missing data. What we
want is to impute the missing data with EM algorithm.
We found Amelia II on Inside R website and tried Amelia II. However, we are
confused with the return value of *amelia {Amelia}. *Will it return the
imputed data or something else? If it does not return the imputed dataset,
what is the next step for us to impute our missing data in the original
dataset so that we could proceed to the next step.
Thanks for your help and patience.
Sincerely,
Yunyan Zhang (Jackie) 张云艳
Industrial and Operations Engineering, B.S.E. 2014
College of Engineering, University of Michigan
Electrical and Computer Engineering, B.S.E. 2014
UM-SJTU Joint Institute, Shanghai Jiao Tong University
jackiezh(a)umich.edu | 734-757-3234
Hello all,
I noticed when trying to run imputations in Amelia 1.7.2 that I
would get error messages indicating some compatibility problems with
the Rcpp package. For instance, I was getting the error:
function 'dataptr' not provided by package 'Rcpp'
I was able to solve this problem by first installing the latest
version of the Rcpp package (0.11.0), and then installing Amelia.
Hopefully this will be helpful to other people having this issue.
Could Amelia be calling for the wrong version of Rcpp? There was a
lot of recent discussion about fixing some compatibility issues
between Rcpp and many CRAN packages (of which Amelia was one), and I
think all the fixes are incorporated into Rcpp 0.11.0.
-Garrett