SDM decision tree to choose best method based on context

[jmlord]

Please add criterias in the discussion below.

[tpoisot]

I'm curious to see what it would look like - maybe @frousseu can share a first draft? I'm sure @gottacatchenall would have ideas about it too.

[frousseu]

We need to start thinking more seriously about this, but to be honest I don't have a very clear idea about what it would look like yet. Assuming we only have presence-only data and a way to account for sampling bias, it comes down to is there a best method or is an ensemble method better? For the amount of data needed to model in the first place, I really like the idea of looking at stabilization of coefficients (https://doi.org/10.1016/j.ecolmodel.2020.109283), predicted values or surface area covered with the amount of data.