Research

Within Criminology, the analysis of criminal behaviour is a central issue. Such an analysis involves different types of aspects, such as biological, psychological and social aspects and their mutual interactions. Usually such analysis are made by criminologist researchers or practitioners, without using any formalisation or computer support. In recent years much progress has been made in biological, cognitive and social complex dynamical systems modelling within areas such as Artificial Intelligence. The methods developed in these areas, when integrated, open the perspective to address the analysis of criminal behaviour in more exact, formalised and computer supported manners. Thus, the way is opened to a more solid basis and computer support for simulation and analysis. In my work this potential is explored. It identifies on the one hand useful knowledge from the literature in Criminology and the different disciplines underlying it, and on the other hand a number of dedicated modelling techniques from areas such as Artificial Intelligence and Cognitive Science. By combining these, integrated computer supported methods to criminal behaviour analysis can be created.

Currently I am leading the Vidi project ‘Integrating sentiment analysis in real crowd management’. Within this project, we will develop an agent-based computational system to predict the dynamics of crowds, improving upon existing models by including individual and collective mental states. By automated analysis of social media feed and using this as input for the simulation model, we will be the first to be able to predict crowd dynamics in real-time, both with respect to behaviour and emotional states like aggression or panic. Eventually, this system can be used offline as basis for virtual training to enable security personnel to practice crowd management procedures and online to support security personnel in managing crowds during events.