Dear fellow researchers,
I have currently faced an error in R when running maxent which indicated; Error in .local(x, p, ...) :
more than half of the absence points have NA predictor values. In any case, the code i used to create my background points were :
#loading bias file and creating background points (10000)
bias_buff
It looks like you want to remove the NAs. There are 2 ways to do that. 1. Delete those observations or 2. Impute the NA values by one R's procedures for doing that. What you do depends on your goal and/or research question. Best, D. Booth
Dear Noh,
I think that this was implemented in package dismo just as a security restriction to avoid interpreting the weak model.
If you are sure that the resulted model will be good enouh for you and you won't misinterpret it, then you can simply remove lines 203:209 from the source code of maxent.R (https://github.com/cran/dismo/blob/master/R/maxent.R), and mention this modification in the publication where you cite dismo/MaxEnt.
If you are unsure, you can omit or impute NA values, as David Eugene Booth suggested, or use more pseudoabsence points from the region where the predictors ar known.
HTH,
Ákos
Thank you so much for the prompt response Ákos Bede-Fazekas David Eugene Booth . Managed to figure out that the extent of my raster file which i used to draw out my psuedo-absences were not in the same extent as the larger than the raster layer files of my environmental variables which then led to a huge portion of the points being in areas where they shouldn't be! Sorry for the confusion!
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology