prior with multiple covariates doesn't seem to work

[acannistra]

From the documentation, we see that it is possible to provide a prior estimate of presence probability given a vector of environmental covariates, but trying this results in the following error:

Error in while (obj.old - obj > tol & it < itmax) { : 
  missing value where TRUE/FALSE needed

You can see our code (line 185) here.

The package works well otherwise, using priors on single environmental variables.

Any advice is appreciated.

[goldingn]

Hi Tony.

I'm a bit confused by your function. The prior argument has to be a function that takes a dataframe (of all covariates) and returns a vector of probability of presence for each row in the dataframe.

You use cbind, which returns either a matrix or dataframe, right?

Sorry if I missed something.

[acannistra]

Oh, I see. For example, if we have environmental covariates A, B, and C, the prior function produces a single probability of presence given the three covariates, rather than a probability given each of the covariates individually? Would it be possible for you to give an example?

[goldingn]

That's correct. It's an a priori guess at the SDM function. There's an example here: https://github.com/goldingn/gp_sdm_paper/blob/master/figures/fig1_2.R#L63

[acannistra]

Ok, great. Thanks for your help!

[goldingn]

No problem!