Pointwise log-likelihood for Mark-Recapture Model

Modeling
, ,
2

I think I found a solution after some digging on this page. Apologies for not finding it before posting a duplicate question. The answer was supplied by user mbjoseph, thanks!

generated quantities{  
vector[nind] log_lik;
for (i in 1:nind) {
    log_lik[i] = 0;
    if (first[i] > 0) {
      for (t in (first[i] + 1):last[i]) {
        log_lik[i] = log_lik[i] + bernoulli_lpmf(1|phi[i, t - 1]);
        log_lik[i] = log_lik[i] + bernoulli_lpmf(y[i, t]|p[i, t - 1]);
      }
      log_lik[i] = log_lik[i] + bernoulli_lpmf(1|chi[i, last[i]]);
    }
  }
}

Thanks,
Jason

2 Likes