search

statnet / ergm.ego

Fit, Simulate and Diagnose Exponential-Family Random Graph Models to Egocentrically Sampled Network Data https://statnet.org
Other
14 stars 4 forks source link

Additional ergm terms abbsdiffby #38

Closed dth2 closed 3 years ago

dth2 commented 4 years ago

Does ergm.ego need custom versions of each ergm term? Sam has written up an ergm term, absdiffby, for targeting age homophily by sex, but when I try to use it with an ergm.ego call I get:

Error in EgoStat.absdiffby(egodata = structure(list(egos = structure(list( : could not find function "EgoStat.absdiffby"

I think this a going to be a fairly standard term used for modeling heterosexuals so it would be helpful if it worked with ergm.ego.

krivit commented 4 years ago

Yes, it needs an EgoStat for every term. I am looking at some algorithmic improvements that eliminate this need, but it'll be a while before I'll have time to work on this.

krivit commented 3 years ago

@dth2 , @martinamorris , do we want to get this into the 3.6 release?

sgoodreau commented 3 years ago

@smjenness has added this term into EpiModel, and my understanding is that it uses different arguments than the version we were planning for ergm (and which I had in evonet). So if we do add it as it, we might create additional downstream issues. @smjenness can you clarify? One option may be simply to give it a slightly different name.

dth2 commented 3 years ago

Yes I think we need this if possible.

Get Outlook for Androidhttps://aka.ms/ghei36

From: Steven Goodreau notifications@github.com Sent: Friday, October 16, 2020 5:47:24 PM To: statnet/ergm.ego ergm.ego@noreply.github.com Cc: Deven T. Hamilton dth2@uw.edu; Mention mention@noreply.github.com Subject: Re: [statnet/ergm.ego] Additional ergm terms abbsdiffby (#38)

@smjennesshttps://github.com/smjenness has added this term into EpiModel, and my understanding is that it uses different arguments than the version we were planning for ergm (and which I had in evonet). So if we do add it as it, we might create additional downstream issues. @smjennesshttps://github.com/smjenness can you clarify? One option may be simply to give it a slightly different name.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/statnet/ergm.ego/issues/38#issuecomment-710712707, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ACP36MPAVUAHNOIDI5BRRQLSLDSRZANCNFSM4JQV5KQQ.

krivit commented 3 years ago

Ah, I thought that it was already in ergm. If it's not, then it shouldn't be in ergm.ego, either. The EgoStat should then be implemented in EpiModel.

sgoodreau commented 3 years ago

maybe I have it wrong and it's in ergm, not Epimodel.

martinamorris commented 3 years ago

It's not in ergm.

@krivit, will it be possible to use the operator terms in 4.0 to produce an absdiff by attribute (or mixingmatrix by attribute)? i confess i still don't fully understand this new interface.

krivit commented 3 years ago

I think so, by interacting absdiff with appropriate nodal or dyadic covariates or filtering on them. Though, we need to think through operator EgoStats as well.

krivit commented 3 years ago

In any case, the term probably shouldn't be in ergm.ego but closer to where its InitErgmTerm counterpart is defined, so I am going to close this.