Introduction to the basic concepts of Bayesian statistics

[2 hours] - [English]

Do you already heard about Bayesian statistics and want to know more about it?

Bayesian statistics is an alternative approach to the frequentist inference (generally taught in basic statistical courses). More and more used by the scientist community, Bayesian statistics enables, among others, to include prior knowledge (through the prior distribution) in the model specification.

Training aims
The conceptual differences between the frequentist and the Bayesian inferential approaches will be highlighted in this seminar. The fundamental notions of Bayesian statistics (prior and posterior distributions of a parameter, among others) will be presented. What do we do when frequentist inference suggests to reject the null hypothesis while the opposite is favored by the Bayesian approach? This issue called the Lindley's paradox will also be discussed.

Prerequisites
This training requires a basic knowledge of statistics.

Content

• Some reminders of inferential statistics
• Comparison between the frequentist and Bayesian approaches (estimation & hypothesis testing)
• Bayes factor
• Lindley's paradox

Methods and method families discussed
Bayesian statistics