MODULE LEADER: Dr Andre Python
EMAIL ADDRESS: andre.python@bdi.ox.ac.uk
DAYS: November 4-5 2019
DURATION: 1.5 days that include lecture and practical sessions. Typically, students should be able to complete the practical in class. However, they are expected to spend 1-2 hours going through the lecture materials before the second day.
AVAILABLE TO: GMS DPhil students (compulsory) and all members (invited) affiliated with the Wellcome Centre for Human Genetics.
MODULE AIM: The aim of the module is to ensure that the students familiarise with the principles of Bayesian statistics and of Bayesian modelling and that they are able to apply such techniques on real data problems.
OVERALL MODULE LEARNING OUTCOMES: By the end of this module, students should be able to: • Compare and contrast Bayesian versus frequentist interpretations of probability. • Understand and explain what a posterior distribution is. • Use R to build Bayesian models and perform Bayesian inference. • Describe the necessary tools for model checking and comparison and be able to implement those in R. • Describe the basics of simple parameter estimation procedures.
DESCRIPTION: This course is compulsory for GMS DPhil students. It introduces the basic concepts related to Bayesian statistics and Bayesian modelling. The skills acquired during the course are applicable to: (i) data analysis and PhD dissertation (for those who choose to carry out Bayesian analysis); (ii) the interpretation of genomics/biomedical papers, which use Bayesian methods for the statistical analyses. In addition, students interested in a career in Biostatistics will particularly benefit from this module.
BACKGROUND KNOWLEDGE REQUIRED: GMS module: Introduction to Statistics for Genomics
Software: R 3.6 and R studio 1.2
Statistics: basic knowledge in probability theory and classical (frequentist) statistics
MODULE STRUCTURE: The course contains a mixture of frontal lectures and practicals. During the lectures the principles of Bayesian thinking will be presented together with examples of their application in social science and epidemiology (regression models, hierarchical models, etc.). Applications of Bayesian models in R will also be presented during the lectures. During the practicals, the students, under the supervision of tutors (and/or the teacher), will have the opportunity to carry out data analysis in R using examples drawn mainly from the social and health sciences.
FILES: Introduction to Bayesian statitsics.pptx: the lecture will start with this power point which will be completed by the students during the course.
Introduction to Bayesian statistics_teacher.pdf: completed pdf (teacher version) of Introduction to Bayesian statitsics.pptx.
practical_nb.1.pptx: this power point will be used to introduce practical 1 (in R)
practical_nb.1.R: the R file corresponding to practical 1
practical_nb.1_answers.R: the R file corresponding to practical 1 with answers
ASSESSMENT: There is no formal assessment for this course. However, an active participation is desired.
FORMATIVE FEEDBACK: Practical sessions always offer great opportunity for receiving formative feedback from the tutors – weekly office hours will be available to those with further questions and remarks. Brief group discussions will be incorporated in some of the lectures to further evaluate students’ progress and insight.
Chapter 2 of Statistical Rethinking, A Bayesian Course with Examples in R and Stan, R. McElreath, CRC Press, 2016
4 November 9:30-12:00
5 November 9:30-16:00