Hi all,
please advise on a statistically sound modelling approach (preferably in R with literature/web references) to use that accounts for many zero observations in a bird abundance x sample site dataset. I want to relate these bird abundances to environmental variables at those sites.
If possible I would like to incorporate a temporal scale by dividing the dataset to three equally-spaced sampling periods for which I have data.
Thanks in advance and please ask if you require more info.
HI Rion,
as i suggested recently in another question you could try to use a Zero Inflated Poisson model, or something similar. Take a look to the ‘pscl’ package in R. it offers some different models. particularly to the ZIPM they should allow you to study the bionomial component and the count component on your data. it look they could fit your scope...
Btw, I am not really in this models, so i dont know if they have a predict function, for istance to do niche modelling or prediction... probably they do have it...
cheers
lorenzo
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling