Bayesian Estimation of Deer Population Dynamics Using Hamiltonian Monte Carlo Algorithm

抄録

For the estimation of deer population dynamics, Bayesian method using Gibbs sampler which is one of the Markov Chain Monte Carlo (MCMC) method has been adopted in the previous studies although the method takes much time for calculation. In this study, we introduce Hamiltonian Monte Carlo (HMC) method as a new approach which is more efficient than Gibbs sampling. To verify our proposal method contributes to deer population managements, we modified the model of a previous study to apply Hamiltonian dynamics and conducted numerical experiments by comparing the results which are estimated in two methods with using simulation based data. As a result, we confirmed HMC method is more time-efficient than Gibbs sampler because the number of samples abandoned for after procedures in HMC method was much less than in Gibbs sampler. Therefore, we conclude that the estimation using HMC is valuable for the estimation of deer population dynamics.