Workshop: Bayesian modelling for observational data

Wednesday, April 17, 2024, 13:00 - 17:00 h, AND 4.55/57, Andreasstrasse 15, 8050 Zurich

In recent years, Bayesian methods have become increasingly popular not only for analyzing experimental data but also as a viable and convenient alternative to frequentist methods in the analysis of observational data. In this workshop, Dr. Gidon Frischkorn, an expert in Bayesian modeling and Monte Carlo simulations will outline the differences between Bayesian and frequentist approaches, how to incorporate prior knowledge, posterior distributions, and the advantages of credibility intervals. Furthermore, attendees will learn, through hands-on exercises, how to use Bayesian methods in R using the “brms” package, how to overcome implementation challenges with Markov Chain Monte Carlo simulations, and how to interpret results effectively. This workshop aims not only to enhance statistical proficiency but also to offer an alternative view and methodology for analyzing observational data.

Program

13:00 - 13:50     Introduction: Bayes Rule & Basic Logic of Bayesian Statistics
13:50 - 14:20     MCMC Sampling: Why is it necessary?
Coffee Break
14:40 - 15:20     Generalized Linear Models in brms
15:20 - 15:50     Statistical Inference with Bayesian Models
Coffee Break
16:10 - 17:00     Practical excersizes: Working with own data or example data

Preparation

This workshop includes practical exercises. So, please bring your own laptop and have STAN and the R-package 'brms' installed prior to the workshop. Instructions for installing STAN (that is the Rpackage ‘rstan’) and the 'brms' package can be found here: RStan Getting Started · stan-dev/rstan Wiki (github.com) and here: Bayesian Regression Models using Stan • brms (paul-buerkner.github.io)

If you encounter any difficulties during the installation process, you can email Lukas Eggenberger, and we will try to help you out as good as we can.

Sign-up Link

Contact details
Gidon Frischkorn (Office BIN 4-B_04)