Benchmark different termination criteria (confidence, Hummel, etc.) and MCMC parameters (fixed vs ESS)

[krivit]

As of this writing, the default is to use "confidence" criterion, but it might be worth going back to "Hummel", and others might be worth trying as well. This choice is controlled by MCMLE.termination= control parameter.

For the MCMC parameters, the default is to use adaptive MCMC with sample size a multiple of the number of free parameters and target ESS starting at 64.

[mbojan]

Try different values of MCMC.termination:

• "confidence"
• "Hummel"
• ...

Using faux.magnolia.high data converted to an egor call ergm.ego( dofit=FALSE ) . This will construct an ergm() call that can be used in testing with different values of control=control.ergm(MCMLE.termination=).