Duration
24h Th, 24h Pr
Number of credits
| Master in environmental science and management (120 ECTS) | 4 crédits | |||
| Master in environmental science and management (60 ECTS) | 4 crédits |
Lecturer
Coordinator
Language(s) of instruction
French language
Organisation and examination
Teaching in the first semester, review in January
Schedule
Units courses prerequisite and corequisite
Prerequisite or corequisite units are presented within each program
Learning unit contents
This course is an introduction to Bayesian Hierarchical Modelling where a model appears as a set of stochastic nodes connected by arrows expressing a causal link between the node « parent » and the node « child « . It is the creative part where the student demonstrated imagination to mimic nature which has produced the available sample. This network can be seen as a series of instructions producing data similar to available observations. Inference or technical part consists to move from observed effects (observations) to causes that produced them (the unknown parameters) using Bayes rule. The intensive use of the methods of Monte Carlo Markov Chain (MCMC) liberates student from most of technical difficulties as much as we use WinBUGS a free software linked to R by the package R2WinBUGS.
Learning outcomes of the learning unit
Train the student to the production of information relevant from a sample of raw data (measures or field observations) in order to help decision maker. In this perspective, the quantification of uncertainties is essential.
Prerequisite knowledge and skills
The student is supposed have basic knowledge in mathematics (secondary level), in statistics (Bachelor level) and he knows the standard laws (Bernoulli, binomial, Poisson, normal) and he is able to manipulate a computer and he can write a simple code.
Planned learning activities and teaching methods
The course is a succession of theoretical courses followed by practical work under the leadership of the Professor, and then in the form of exercises to perform alone or group depending on the case.
Mode of delivery (face-to-face ; distance-learning)
Face-to-face for theoretical courses and some practical works
Recommended or required readings
Pratique du calcul bayésien (Boreux et al., 2010)
Assessment methods and criteria
The control of knowledge consists of a problem that the student performs on his laptop during a limited time (typically three hours) and at the end of which he sends a file saved as Professor email address. The exam is open book and the evaluation criteria exclude therefore the reproduction of knowledge to focus on creativity, common sense and critical analysis of the results obtained.
Work placement(s)
None.
Organizational remarks
None.