Eric,
Another thing you can try, though I bet this won't fix the problem,
would be to load the subset of the data which only contains the four
variables you are using.
Kosuke
---------------------------------------------------------
Kosuke Imai Office: Corwin Hall 041
Assistant Professor Phone: 609-258-6601 (Direct)
Department of Politics eFax: 973-556-1929 (Direct)
Princeton University Email: kimai(a)Princeton.Edu
Princeton, NJ 08544-1012
---------------------------------------------------------
On Fri, 21 May 2004, Olivia Lau wrote:
Dear Eric,
When reporting problems, you should include the traceback(). As you may
know, R is memory-inefficient. As mentioned in the Zelig manual,
http://gking.harvard.edu/zelig/docs/If_You_Are.html
you should seriously consider running your enormous model on a server.
I've successfully run a R logit model with up to 1 million observations on
a small Linux box, so 500,000 observations with only 3 covariates should
be fine. The advantage of linux R over windows R is precisely that linux
R is much more efficient.
If you insist on running windows R, try running your model using the
following syntax:
z.out <- glm(mzcowwar ~ territor + cc24 + cc24xt, data = maoz,
family = binomial(logit))
If you get the same error as before (which I am pretty sure you will), you
want to email the r-help list with the synatx described above, not the
syntax you have below. This is a problem with glm() not zelig().
In addition, you want to include the following information with your
r-help email:
dim(maoz)
and your platform (windows, operating system, processor and speed),
version of R (windows R 1.9.0, I suspect), etc.
Yours,
Olivia Lau
Zelig[[3]]
On Fri, 21 May 2004 emwilk(a)buffalo.edu wrote:
I've run into a problem running analysis on
a rather large (nearly
500,000 cases) data set. When I attempt to run the analysis, I get the
following error message:
library(foreign)
wd<-"C:/Documents and Settings/PSGSA/My Documents/Eric/maoz"
setwd(wd)
maoz<-read.dta("maoz.dta")
z.out<-zelig(mzcowwar ~ territor + cc24 + cc24xt, model = "logit",
data = maoz)
Error: cannot allocate vector of size 3900 Kb
I've already increased the memory and vector allowance (as outlined in
the F.A.Q. and message board) using the --max-vsize command but still
have had no luck. I'm wondering if this is due to the memory of my
machine or to R. I was able to run the same procedure with the same
variables and data set in Stata so I'm guessing it's not completely my
computer but rather that I'm doing something wrong in Zelig. If anyone
could give me a suggestion, it would be greatly appreciated.
Sincerely,
Eric
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)latte.harvard.edu
[un]subscribe Options:
http://lists.hmdc.harvard.edu/?info=zelig
Zelig program information:
http://gking.harvard.edu/zelig/
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)latte.harvard.edu
[un]subscribe Options:
http://lists.hmdc.harvard.edu/?info=zelig
Zelig program information:
http://gking.harvard.edu/zelig/
-
Zelig Mailing List, served by Harvard-MIT Data Center
Send messages: zelig(a)latte.harvard.edu
[un]subscribe Options: