[Rsiena-help] crash during estimation of bipartite behaviour model

[Nate Doogan]

Hi Marco.

I'll throw out a guess. The dfra object is probably the initial derivative
estimate. If a value in the matrix is missing, I guess it means the partial
derivative could not be estimated. Maybe this is because of something funny
in your covariate coding. Perhaps the first thing to do is look for errors
in your data. I could also imagine that a lack of variability in a
covariate could cause some issues.

Good luck. I hope some more informed guesses will emerge.

Nathan Doogan


On Sun, Nov 27, 2011 at 10:24 PM, marco tonellato <m.tonellato at gmail.com>wrote:

> Dear All,
>
> while running siena07 (RSiena rev=184 on a multi-core Linux-based machine)
> to estimate a coevolution network/behaviour model using bipartite networks,
> I encountered the following problem:
>
> I specified the model with a dependent bipartite network (4 sparse
> matrices, 135x719 nodes), a bunch of strucural (cycle4, outActSqrt,
> inPopSqrt, outInAss) and individual covariates, a dependent behavioural
> variable (related to the first nodeset), and a bunch of behavioural
> parameters (avSim and outdegree, plus of course the defaults linear and
> quadratic shape). Then I launched siena07 and I was able to get through
> phase1, until this error message popped up:
>
> Error in if (dfra[i, i] < 0) { : missing value where TRUE/FALSE needed
>
> Can you please help me on this?
>
> Thanks a lot.
> Marco
>
>
> _______________________________________________
> Rsiena-help mailing list
> Rsiena-help at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rsiena-help
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rsiena-help/attachments/20111127/83f27798/attachment.htm>