ENMTools
ENMTools copied to clipboard
Build in bootstrap and jackknife functions
We should set it up to do bootstrap and jackknife over the data to estimate sampling variance, but also to do variable jackknifes for percent contribution.
For now here's an Rmd file showing how to do bootstraps and repeated train/test partitioning manually, if anyone needs to.
Building-replicate-models.pdf
I'm thinking we might want to generalize the rts.rep code to do this. We can basically just have a "reps type" and "nreps", and then have RTS, the Bohl et al. test, the ENMTools version (RTS on test data), bootstrap, and jackknife. Possibly even allow passing in a vector to do more than one in a single go?