14 Jul
14 Jul
6:25 a.m.
Hi Chunling,
There is some suggestion that imputing on more than 20% of missing data becomes
problematic. The second issue is the nature of the missing data, are they missing
completely at random, at random, or due to some systematic effect. This also has an
impact on missing data imputation.
Thanks Paul
----- Original Message -----
From: Lu, Dr Chunling
To: amelia(a)lists.gking.harvard.edu
Sent: Monday, July 14, 2008 3:57 PM
Subject: [amelia] a general question about using Amelia
Dear friends,
Amelia is used to impute missing values, but does the percentage of missing affect the
imputation (I assume so), is there any cutoff point for "good imputation" or
"warning"? I have a cross-country data with 43% of them are missing. Is Amelia a
good choice?
Thanks.
Chunling
- Amelia mailing list served by Harvard-MIT Data Center [Un]Subscribe/View Archive:
http://lists.gking.harvard.edu/?info=amelia
15 Jul
15 Jul
12:28 p.m.
New subject: Can't process imputation
Dear friends,
We are trying the Amelia (version 1.1-29) command in R (version 2.7.0)
with the following code
data_output <- amelia(data=data_input,m = 10,outname = "who_im",empri =
10, ts="year", cs="country1",
lgstc=c("gge.gdp.wb","ex.gdp"),p2s=2)
The program always stuck at the following stage:
amelia starting
beginning prep functions
running bootstrap
Any ideas?
Thanks very much.
Chunling
5771
days inactive
5772
days old
amelia@lists.gking.harvard.edu
3 comments
3 participants
participants (3)
-
Gary King -
Lu, Dr Chunling -
paulandpen