Implement more complex durational terms

[krivit]

So far, only mean.age is implemented.

What terms are needed?

[skyebend]

• periodic terms (cyclic tie probability as a function of model time, to capture weekly/seasonal activity patterns)
• as a function of categorical variables on edges or vertex dyads. I notice one of steve's models is keeping two seperate network objects, one for long duration ties and one for short duration ties. Seems like ideally should be able to express this in a single model I forget what some of the other ones were, maybe they will come back

[krivit]

Thanks, Skye.

Edit: My bad, didn't realize this was for ergm.ego.

periodic terms (cyclic tie probability as a function of model time, to capture weekly/seasonal activity patterns)

I don't know what the "sufficient statistic" for that would be. Any thoughts?

as a function of categorical variables on edges or vertex dyads.

Mean age by covariate should be doable.

I notice one of steve's models is keeping two seperate network objects, one for long duration ties and one for short duration ties. Seems like ideally should be able to express this in a single model

That's a whole separate project...

[skyebend]

Oh, I thought it was for tergm, no idea about durational terms for ergm ego