Bayesian Modeling of AgeâSpecific Survival in Nesting Studies Under Dirichlet Priors
2001
He, Chong Z. | Sun, Dongchu | Tra, Yolande
There has been much work done in nest survival analysis using the maximum likelihood (ML) method. The ML method suffers from the instability of numerical calculations when models having a large number of unknown parameters are used. A Bayesian approach of model fitting is developed to estimate ageâspecific survival rates for nesting studies using a large class of prior distributions. The computation is done by Gibbs sampling. Some latent variables are introduced to simplify the full conditional distributions. The method is illustrated using both a real and a simulated data set. Results indicate that Bayesian analysis provides stable and accurate estimates of nest survival rates.
显示更多 [+] 显示较少 [-]