Maximum entropy in Suitability Modeler?

[sdegroot]

Just curious if an algorithm like maximum entropy or a Bayesian classification might be implemented in the suitability modeler. While it's very helpful to have charts to immediately see the effects of your classification or weighting or function choice, many times I have a poor idea what features of the habitat are most important, making it hard to assign variables when reclassifying or weighting layers. With some point locations (which often do not represent the full range of a species), Maxent ( https://biodiversityinformatics.amnh.org/open_source/maxent/  ) seems to do a decent job of returning reasonable areas of suitable habitat, even if I throw in 80 rasters and let it sort out which are most important. I would enjoy seeing this as choice in the modeler, just so I could see what affect it has on the final output. In some cases, it might not be the best model, but right now I have no way of knowing.

It would also be interesting if the model could provide a Bayesian posterior probability or some other measure to show how well it fits our input data (i.e. the probability of the model, given our data). Perhaps at each step, to help us decide which value or function is most appropriate?

See also Maxent algorithm added to ArcMAP