ch06

Chapter 6: likelihood analysis of spatial capture-recapture models

These models use a grid approximation to marginalize over activity centers s (as you would do for maximum likelihood estimation), but are Bayesian. This marginalization is not required in Stan (which can directly use s), but these models show how it can be done.

Known N (SCR0)

See scr0-known-n.R.

Unknown N (SCR0)

1. Binomial form. See scr0-unknown-n-binomial-form.R, which uses a binomial model for N (as in eq. 6.2.1).

2. Data augmentation. See scr0-unknown-n-data-augmentation.R.

Wolverine case study

1. Binomial form. See scr0-wolverine.R.

2. Poisson integrated. See scr0-wolverine-poisson-integrated.R. Notice that even though we integrated s out of the model, we can still sample from s in the generated quantities block.