Iker Leanizbarrutia
University of the Basque Country

Abstract
Markov Chain Monte Carlo (MCMC) is the best way to gather information and constrains of parameters from observational data within the Bayesian approach. This mathematical tool is very used in cosmology, where we only have a single realisation of the universe. I will explain the theoretical background of the MCMC as well as its general features, and I will give details of the Metropolis-Hastings algorithm. As an example, I will show the results of MCMC applied to a UDM model with fast transition (PhysRevD.93.043537), explaining the process with certain details regarding MCMC and model comparison.

2016 April 08, 14:00

IA/U.Lisboa
Faculdade de Ciências da Universidade de Lisboa (C6.2.51)
Campo Grande, 1749-016 Lisboa