• interaction patterns
• mobility interaction
• moving agents
• pattern mining
• simulation

In this thesis we propose an analysis framework for studying the interactions between moving objects, such as cars on roads, pedestrians in a square, etc. We follow a data mining approach, based on the computation of simple interaction events and on the extraction of complex patterns, describing frequent combinations of events that happen together and their evolution in time. The work includes two case studies on real datasets, respectively on cars and roads.