The course, which will be held at ICES Secretariat November 27 – December 1, will train participants in the fundamentals of the versatile modelling technique known as the Bayesian method, looking at ways of analysing different research questions including those related to social-cultural systems.
The strengths of Bayesian lies in being able to input and combine different sources of quantitative and qualitative data, factoring in both uncertainty and preferences depending on the question. Working these preferences and uncertainties as well as avoidances into the model then helps present decisions or steps to be taken to manage the system in question. The course will explore some of the more useful ways of doing this.
Students will be taught the basics of Bayesian Network modelling that can be applied both to natural and social science issues. The course will look at problem framing, how to build the model structure, and how to quantify the model. Existing social-cultural approaches will also be looked at, with an emphasis placed on what to consider when, for example, interviewing experts to populate models.
"Bayesian networks are graphical, meaning complex coding is not necessary, and that they are accessible and suitable for a wide range of users and disciplines," explained Laura Uusitalo from the Finnish Environment Centre's Marine Research Centre, who is instructing the course alongside Päivi Haapasaari of the University of Helsinki.
"We welcome students from various scientific backgrounds. If they turn up with a problem in mind then great, we can work on that, If not, we will have some examples they can experiment with."