R help - May 2012 - ergm model, nodematch with diff=T

Dear all,

I am new to network analysis, but since I have good data I started to read
about it and learned how to use the ergm and related packages.  I generally
get interesting results, but when I run a model including sociality and
selective mixing effects for different groups, the model runs (and
converges) but I get a warning as follows:

mod <- ergm(network ~ edges + gwesp(0, fixed=T) +
nodefactor("code") +
nodematch("code", diff=T)+ nodecov("vulnEVI") +
absdiff("vulnEVI"))
Observed statistic(s) nodematch.code.5, nodematch.code.6, and
nodematch.code.8 are at their smallest attainable values. Their coefficients
will be fixed at -Inf.

The model nevertheless runs and converges and produces the following output:
=========================Summary of model fit
=========================
Formula:   network ~ edges + gwesp(0, fixed = T) + nodefactor("code")
+
    nodematch("code", diff = T) + nodecov("vulnEVI") +
absdiff("vulnEVI")

Iterations:  20 

Monte Carlo MLE Results:
                  Estimate Std. Error MCMC % p-value    
edges             -5.34531    0.66468     NA < 1e-04 ***
gwesp.fixed.0      2.17907    0.27527     NA < 1e-04 ***
nodefactor.code.3 -0.19424    0.20055     NA 0.33283    
nodefactor.code.4 -0.45733    0.21730     NA 0.03538 *  
nodefactor.code.5 -0.49081    0.22815     NA 0.03151 *  
nodefactor.code.6 -1.19468    0.25695     NA < 1e-04 ***
nodefactor.code.7 -1.26290    0.24369     NA < 1e-04 ***
nodefactor.code.8 -1.26365    0.38244     NA 0.00096 ***
nodematch.code.2   0.10071    1.31137     NA 0.93879    
nodematch.code.3   2.70398    0.53684     NA < 1e-04 ***
nodematch.code.4   0.67157    0.26456     NA 0.01117 *  
nodematch.code.5      -Inf         NA     NA      NA    
nodematch.code.6      -Inf         NA     NA      NA    
nodematch.code.7   0.58640    0.81906     NA 0.47407    
nodematch.code.8      -Inf         NA     NA      NA    
nodecov.vulnEVI    0.22562    0.07024     NA 0.00133 ** 
absdiff.vulnEVI   -0.01086    0.11925     NA 0.92742    
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1 


Hence the model is unable to obtain estimates for three of the groups and I
don't understand why. When I include the nodematch with diff=F there is no
problem (although of course I only get one coefficient). Does anyone know
what the problem might be? At first I though there might not be any edges
connecting members of groups 5, 6, and 8, but I checked this and they
certainly do cooperate among them. Any help would be greatly appreciated.
Thank you!

Best,
Florian


--
View this message in context:
http://r.789695.n4.nabble.com/ergm-model-nodematch-with-diff-T-tp4620052.html
Sent from the R help mailing list archive at Nabble.com.

Hi all,

I have a question about using bigmemory package.

Here is my code:

 > 
x=read.big.matrix("acc3.dat",backingfile="acc3.bin",descriptorfile="acc3.desc",type="double")
Error in filebacked.big.matrix(nrow = nrow, ncol = ncol, type = type,  :
   A big.matrix must have at least one row and one column

And here is the example code:

 > x <- read.big.matrix("airline.csv", header=TRUE,
+ backingfile="airline.bin",
+ descriptorfile="airline.desc",
+ type="integer")

So, what was wrong? Any suggestions please?

Thank you very much.

ya

Hi all,

In addition, I am planning to do a multiple imputation with MICE package 
using the data read by bigmemory package.

So usually, the multiple imputation code is like this:

 > imp=mice(data.frame,m=50,seed=1234,print=F)

the data.frame is required. How can I change the big.matrix class 
generated by bigmemory package to a data.frame?

Thank you very much.

ya