Complex models tutorial

"Design of Experiments for Complex (and complicated!) Models" course on 14-15/10/2015 at IMF-CSIC, c/Egipciaques, 15 - 08001 Barcelona, Spain.

The SimulPast project is organising, on the 14th and 15th of October 2015, a two-day tutorial course on the analysis of complex simulation models. Dr. María Pereda, a postdoctoral researcher of SimulPast group G10 (University of Burgos) and working on Case Study 3, will be teaching how to apply statistics and machine learning to analyse your models.

Places are limited to 15 participants and these will be allocated on a first-come first-served basis. Members of the SimulPast project are given priority, but the event is open to other people as well. Anyone interested in participating should send an email (with #DOECM as subject) to This email address is being protected from spambots. You need JavaScript enabled to view it. by Friday 9th October 2015.

For the practical part (second day), participants are encouraged to bring their own laptop with R, R Studio and NetLogo installed. We will also provide the necessary installation executables for those who will need them.

For further details, please see the full programme attached.

 

"For the purpose of this course, we will consider that a model has a large number of parameters when they are more than five. It may also happen that it is difficult to establish a direct correspondence between parameters’ values and the reality under study. These situations make the analysis of the influence of the parameters in the model output really complicated.

Using statistical sampling and machine learning techniques can help us to attempt a first approach to the understanding of the implications of the parameters on our model, and to guide the conception of more profound and design-detailed experiments. Moreover, we need to ensure that we are covering the whole parameters’ space in our exploration, which is not always the case with traditional approaches, such as Monte Carlo sampling.

In this course you will learn how to design experiments for a model with a large number of parameters using advanced statistical sampling and to analyse simulations results with machine learning techniques."