On Thursday 3 December 2015 dr. Marie-Colette van Lieshout (CWI and UT) held her inaugural lecture as a full professor of Spatial Stochastics at the Faculty of Electrical Engineering, Mathematics and Computer Science of the University of Twente. Her inaugural speech was titled 'Chances in Space and Time'.
In spatial stochastics – the study of phenomena in space and time where uncertainty plays a role – the influence of explanatory variables (also known as 'covariates') on spatial patterns is studied. For example, in gas extraction in Groningen not only the link between observed patterns (quakes) and covariates (gas extraction) is of interest, but also the structure of the spatial patterns (aftershocks) and patterns over time (heavier quakes).
In this discipline, Van Lieshout earlier contributed to the development of conditionally specified models for point processes, and, more generally, random sets, as well as to a correct understanding of the associated interaction structures. She has also developed Monte Carlo methods, including exact simulation methods. The combination of these moment measures provides powerful statistics for exploratory data analysis and model validation that can be used for non-homogeneous data.
The department of Stochastic Operations Research at the University of Twente, to which Van Lieshout is affiliated as a part-time professor since 1 February 2015, is among the world leaders in the field of stochastic operations research. This discipline focuses on the optimization of processes in which uncertainty plays a role. An integrated theory of space-time point processes, however, is still in its infancy, as Van Lieshout addresses in her inaugural speech. Often, either space or time is taken as a mark. In those cases in which space and time are equally important, the chosen models are generally separable, i.e. the time measurements factorize in space and time. Such an approach is useful, but seems less realistic. Also, prediction deserves further attention and new theories, both in the temporal sense of predicting future behavior, as spatially in the sense of completing a partially observed pattern. The same can be said of optimization and control problems for dynamically evolving patterns. A hierarchical approach including unknown parameters provides perspective here, Van Lieshout said.
At the University of Twente, Van Lieshout is a part-time professor for a day in the week. For four days a week, she is affiliated with the Centrum Wiskunde & Informatica (CWI), since 1946 the national research institute for mathematics and computer science in the Netherlands.
Source: Universiteit Twente
Picture: Marie-Colette van Lieshout (CWI and University of Twente), part of the picture. Source: Rinus Baayens/UT.
See also CWI's earlier news item: http://www.cwi.nl/news/2015/marie-colette-van-lieshout-appointed-professor-in-spatial-stochastics