In the case of long reh's and/or large (sparse) networks, it would be useful to fit a rem using a riskset that only consists of the observed dyads. So if the argument 'omit_dyad' is (for instance) set to "inactive", then all inactive dyads would automatically be omitted in the riskset. An extension that could be interesting is to allow the possibility that is changed for different periods (e.g., because the active network is different in the summer than in the fall). I also posted this in remify.
Btw for the aom I was just wondering whether the omission of certain dyads works ok because it would imply that the size of the risk can be different depending on the sender.
In the case of long reh's and/or large (sparse) networks, it would be useful to fit a rem using a riskset that only consists of the observed dyads. So if the argument 'omit_dyad' is (for instance) set to "inactive", then all inactive dyads would automatically be omitted in the riskset. An extension that could be interesting is to allow the possibility that is changed for different periods (e.g., because the active network is different in the summer than in the fall). I also posted this in remify.
Btw for the aom I was just wondering whether the omission of certain dyads works ok because it would imply that the size of the risk can be different depending on the sender.