Day 1: the prior, the likelihood, and the posterior

Outline

In the first part, we will introduce the basic rationale behind Bayesian statistics. We start by explaining the three key components of a Bayesian model: the prior, the likelihood, and the posterior.

Then, we switch to the estimation of parameters by first introducing the basic idea of grid approximation and then outlining the basic idea of MCMC sampling.

At the end of the session we introduce brms and learn how to estimate a simple regression model with brms and just use the summary() and plot() functions to get insight in the model results.

Materials

Slides

The htlm-version of the slides for this first part can be found here

Data

For this first part, we use a straightforward dataset on predicting racetimes for a marathon. The data can be downloaded here (right-click to save as).

References and resources

Data comes from Kaggle

Paul Bürkner’s presentation available on YouTube: click here

Interactive tool demonstrating MCMC sampling: click here

brms homepage: click here