Individual Based Modelling in Epidemiology: A Practical Introduction 27-31 March 2017, Stellenbosch, South Africa
Prof Wim Delva (SACEMA, Ghent University, Hasselt University and KU Leuven) and Dr Lander Willem (Antwerp University) will be presenting an intensive five-day course on Individual-based modelling in epidemiology, organised by SACEMA. The course will run from 9 am to 4 pm daily, 27-31 March 2017, and be held at STIAS (adjacent to SACEMA) in Stellenbosch.
The Registration Form is available here. Please note:
- The deadline for early registration is 31 January 2017.
- For participants within South Africa, the course fee is R4000 for early bird registration and R4500 for later registration.
- For international participants, the fee is 400 € for early bird registration, and 450 € for late registration.
- The fee includes refreshments, lunches, social events, a copy of the textbook, and a non-refundable registration fee (R850 for South African participants, 85 € for international participants).
The costs of accommodation, breakfast, and dinner are not included. Short-term accommodation is in high demand, so best to book early. Useful websites include Airbnb, TripAdvisor, Sleeping out and booking.com.
Individual-based models (IBMs), also frequently referred to as agent-based models, are a relatively new class of models that can be used to gain insight into the population dynamics of complex systems that emerge from the characteristics and behaviours of individuals in the population. This course aims to give participants the skills to design, implement, and analyse IBMs. First, we will introduce the fundamentals of modelling, as well as a popular, open-access platform for building IBMs (RNetLogo). Next, several key modelling concepts will be discussed, including:
- Deciding on appropriate structure and complexity of the model
- Developing rules for the actions of individual
- Analysing the emerging model dynamics
- Fitting the model to empirical data
- Validating model inputs and outputs
Each of these concepts will be illustrated with hands-on examples that participants can run on their own laptops. In the second half of the course, two “state-of-the-art” individual-based models will be explained, run and analysed in detail. They will serve as examples of full-blown implementations of IBMs, intended to address contemporary research questions. Participants will be encouraged to share and discuss the development and fine-tuning of their own ongoing IBM-based modelling work.
The emphasis of this course will be on the basic concepts behind IBMs and on applying these in simple “toy models”, implemented with RNetLogo. The emphasis will not be on the actual programming of IBMs but rather on the process of designing, testing and analysing IBMs to address complex questions in epidemiology.
- Post-graduate students and health science professionals whose work potentially involves the design and/or use of IBMs in epidemiology.
- Ideally, participants should have used R previously for data analysis and simple programming (e.g. writing your own function). However, this is not an absolute prerequisite. R newbies will be asked to work through selected R tutorials before arriving at the course.
A medical doctor and epidemiologist, Wim Delva has a joint research appointment at SACEMA (Stellenbosch University, South Africa) as well as Hasselt University, Ghent University and KU Leuven in Belgium. He is interested in the application of the statistical, epidemiological and mathematical modelling techniques to describe and analyse the behavioural and biological processes underlying HIV epidemics in sub-Saharan Africa and Europe. His current research centres around the inference of sexual network structure, using phylogenetic tree data and behavioural survey data. Other ongoing research projects seek to explore the role of age-mixing patterns and heritability of HIV set point viral load in HIV transmission dynamics, as well as the impact of biomedical and behavioural interventions on HIV incidence.
Lander Willem (MSc Bio-engineering, interdisciplinary PhD Medical Sciences & Sciences) holds a post-doc position at the University of Antwerp. His research targets agent-based modeling in the field of infectious disease transmission, with a particular focus on model exploration, parameter estimation and social contact patterns. In the philosophy of engaging in interdisciplinary research, he is involved in many projects benefiting from his disease modeling and computational skills.