Study of mobility patterns

Abstract

A trajectory can be observed from a geometric point of view, a curve through space, but also from a semantic point of view, by using background information to interpret its course. This allows us to define semantically enriched trajectories, which are annotated with background information. These semantically enriched trajectories show movement behavior. As any type of behavior shows patterns, it is interesting to find these in movement. One option to analyze trajectories is association analysis, a group of techniques with the purpose of finding rules describing found patterns. Apriori is one of the most important association analysis tools. This algorithm has already been adapted to process sequential data, a class of data models trajectories belong to, resulting in AprioriAll. The goal of this work is to describe an Apriori implementation that is adapted to the specifications of semantically enriched trajectories. Our algorithm focuses on the specific nature of trajectories by defi