I think the thought is that, although we can model simple (1st order) interactions directly, we cannot accurately model complex (higher order) systems directly. So multi-domain models integrating all the simple models can be useful, as long as we have some insight into how the domains interact.
Hi David,
I think the thought is that, although we can model simple (1st order) interactions directly, we cannot accurately model complex (higher order) systems directly. So multi-domain models integrating all the simple models can be useful, as long as we have some insight into how the domains interact.