I’d like to implement a hybrid model using the SciML ecosystem, that switches between a jump process (at ‘low’ population sizes) and an SDE. Individual examples of these models are here and here. Any suggestions on best practices to switch between these two types of model as the integration progresses?
Not yet, but it is a big research project that @isaacsas and I are starting up so 
I’m sure we’ll have more and better examples in the near future. My guess though is that the SDE likely isn’t a good idea and you might just want the ODE and the jump process since that would incorporate the noise when/where it matters, but that’s just a hunch. You can set it up with either.
Thanks @ChrisRackauckas! Glad to hear you and @isaacsas are working on this. In the meantime, are there any existing examples to borrow from?
Just the ones you linked.
This is the one example someone put together using existing components: