Integrated habitat models and missing values problem
Posted: Mon Jan 25, 2016 1:38 pm
I've encountered an odd behavior of PRESENCE (running 10.2 currently) when using the integrated habitat models. For the habitat-related parameters (pi,eta), the intercept and beta "blow up" when I use covariates that have missing data that correspond to missing occupancy data. The estimates, using the same covariates, are fine when estimating occupancy dynamics or detection.
I understand there are problems when occupancy data are not missing, but covariates are, but this is not the case here.
Any advice would be appreciated.
I understand there are problems when occupancy data are not missing, but covariates are, but this is not the case here.
Any advice would be appreciated.