Bayesian modelling over the past decade has become an integral part of statistical modelling in ecological and conservational studies. Given the prior knowledge we have, we can then infer from the available data by accounting for this prior knowledge.
A method well known to the Bayesians is the Markov Chain Monte Carlo. Its primary use is to provide a method for generating a sample from which expectations of functions of a random variable X can be estimated. However, this algorithm tends to run slow.
Another interesting algorithm that can be used to increase the computing time is the Integrated Nested Laplace Approximation (INLA). This is used for the class of latent gaussian models, of which most spatio-temporal models belong.
To extend the INLA methodology to incorporate a wider range of models that cannot be fixed using INLA unless some parameters are fixed, the MCMC + INLA approach have been proposed.
It sounds interesting right? For further details, the reader is referred to the original paper (https://arxiv.org/pdf/1701.07844.pdf). This was the first topic discussed at the Statistics Symposium help at by the Statistics group in NTNU. Updates of the other talks will follow soon.
Written by Kwaku Peprah Adjei; Jorge Sicacha Parada