[R-sig-eco] Categorical variables Maxent dismo

[Bede-Fazekas Ákos]

Dear Irene,

Maybe after one-hot encoding of your multilevel factor (i.e. creating 
several dummy/binary/logical variables from the levels of the factor) 
you can train a MaxEnt model that can answer your question. And also 
partial response curve (dismo::response()) for a factor variable will 
show you some information about the difference of the levels.

HTH,
Ákos Bede-Fazekas

2021.01.15. 11:38 keltezéssel, Irene Rojo írta:
> Hola Diego,
>
> Thank you very much for your reply. I have tried the dropLayer function as
> you suggest but it removes the variable ("Fence" in your example) from the
> raster stack "predictors" with all the variables, so I think it is not
> doing what I wanted. Thanks a lot anyway for your time.
>
> Any other advice would be much appreciated.
>
> Regards,
>
> Irene
>
>
>
> El vie, 15 ene 2021 a las 10:10, Diego GT (<dgt3087 using gmail.com>) escribió:
>
>> Hola Irene
>>
>> I have been learning Maxent this week. In my case I have one categorical
>> variable. I tried the ENMeval package there is an option to name
>> categoricals="your variable". On the dismo package I think you have to
>> pred_nf <- dropLayer(predictors, 'Fence') to do the training and
>> background and then in the model you attached them again.
>> xm <- maxent(predictors, pres_train, factors='Fence')
>>
>> I don't know how it works statistically. But I hope it helps.
>>
>> Cheers
>> Diego
>>
>>
>> On Fri, 15 Jan 2021 at 19:33, Irene Rojo <ire.rojo using gmail.com> wrote:
>>
>>> Dear all,
>>>
>>> I am running Maxent models with dismo package. I have several categorical
>>> variables, which have been specified in the model as factors, but the
>>> output gives me only the contribution of the whole variable and not the
>>> category. Is there a way that the analysis reports the category of the
>>> variable which contributes the most? I have not been able to find how to
>>> do
>>> it.
>>>
>>> Thanks a lot,
>>>
>>> Irene
>>>
>>>          [[alternative HTML version deleted]]
>>>
>>> _______________________________________________
>>> R-sig-ecology mailing list
>>> R-sig-ecology using r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>>>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology