When a two-mode REH is supplied (probably easiest by allowing an argument for this, although the algorithm could detect it automatically), the riskset should automatically be restricted to inter-mode relations. (as I understand it, this is already standard in goldfish).
This restriction in the riskset should be maintained throughout the modeling pipeline, although the researcher might want to add some additional restrictions to the riskset as well.
For a two-mode REH, perhaps we should make sure all senders are of one type and all receivers are of the other. That way, the user need not specify which vertex belongs to which type.
When a two-mode REH is supplied (probably easiest by allowing an argument for this, although the algorithm could detect it automatically), the riskset should automatically be restricted to inter-mode relations. (as I understand it, this is already standard in goldfish). This restriction in the riskset should be maintained throughout the modeling pipeline, although the researcher might want to add some additional restrictions to the riskset as well.
For a two-mode REH, perhaps we should make sure all senders are of one type and all receivers are of the other. That way, the user need not specify which vertex belongs to which type.