Graduate Student Seminar I Supranta Sarma Boruah, Bayesian data analysis with Markov Chain Monte Carlo (MCMC) methods

MS Teams

Speaker

Supranta Sarma Boruah | Applied Math, University of Waterloo

Title

Bayesian data analysis with Markov Chain Monte Carlo (MCMC) methods 

Abstract

Bayesian data analysis techniques have found a widespread applications in the scientific fields over the last few decades. In particular. Markov Chain Monte Carlo methods are widely used for statistical inference. 

In this seminar series, I will introduce the concepts of Bayesian data analysis, in particular, using MCMC methods. I will start with the familiar example of linear regression, interpreted in the Bayesian language and then graduate to examples with real cosmological data. In the second part of the series, I will introduce MCMC algorithms beyond the usual Metropolis-Hastings algorithm, which are beginning to be used more extensively and diagnostic tests useful for analyzing properties of the MCMC runs. Finally, in the third part of the series, I will introduce the concept of Bayesian hierarchical modelling, which through its probabilistic modelling captures known relations between different variables. 

The sessions will be a hands-on. Bringing laptops with jupyter notebook installed is encouraged.