[Rsiena-help] Accessing the simulated networks
Brian Rubineau
brubineau at cornell.edu
Tue Mar 8 20:46:29 CET 2011
Dear RSiena experts,
I'm trying to use the simulation engine in RSiena for fixed network models.
That is, the output I'm interested in is the networks in the $sims
portion of an
RSiena estimation object, not the estimation results itself. My problem is
that I'm only getting "NULL" in the $sims list.
I'm using R 2.12.0 and RSiena 1.0.11.119.
Here's a simple example of what I'm trying to do, and the problem:
library(statnet) ## to get the random graph function rgraph
two <- rgraph(20,m=2,tprob=0.2)
two[1,,] <- 0 ## set the 1st graph to empty
two[1,sample(1:20,1),sample(1:20,1)]<-1 ## make the 1st graph have only
1 random tie
two.siena <- sienaNet(array(c(two[1,,],two[2,,]),dim=c(20,20,2)))
two.siena.data <- sienaDataCreate(two.siena)
two.siena.model <- sienaModelCreate(useStdInits = FALSE, projname =
'two',n3=10)
two.effects <- getEffects(two.siena.data)
two.effects$fix <- two.effects$include ## to fix all the parameters
two.1 <- siena07(two.siena.model,data=two.siena.data,effects=two.effects)
two.1$sims
[[1]]
NULL
[[2]]
NULL
[[3]]
NULL
[[4]]
NULL
[[5]]
NULL
The errors from the siena07 process are:
> Error in solve.default(z$msfc) :
> Lapack routine dgesv: system is exactly singular
> In addition: Warning message:
> In max(abs(tstat)[!z$fixed & !z$BasicRateFunction & z$resist > 0.9]) :
> no non-missing arguments to max; returning -Inf
I'm attaching the out file from this process.
Any ideas as to what I'm doing wrong?
Thanks so much!
Best,
Brian
--
Brian Rubineau
Assistant Professor, Organizational Behavior 394 Ives Hall
School of Industrial and Labor Relations ph. 607.255.3048
Cornell University fax. 607.255.2261
Ithaca, NY 14853 brubineau at cornell.edu
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: two.out
URL: <http://lists.r-forge.r-project.org/pipermail/rsiena-help/attachments/20110308/f9028ee7/attachment.asc>
More information about the Rsiena-help
mailing list