Προστασία ιδιωτικότητας σημασιολογικών δεδομένων κίνησης με χρήση μηχανισμού-ελέγχου ερωτημάτων
Privacy preservation of semantic trajectory data with query auditing-mechanism
The use of smartphones and many other devices that support positioning systems and software is an integral part of our everyday life. The increase in their use and application is rapid, resulting in a huge and ever increasing volume of data recorded as moving objects trajectories. The exploitation of recorded traffic data, combined with the modern data mining techniques flourishing, was a natural consequence from both the researchers in the survey process and the providers’ organizations mostly for business reasons. However storage and analysis of geospatial data not only offers innovations and knowledge, but also introduced the risk of disclosing private information of the recorded entities to malicious users. This fact has resulted the development of theories and techniques with the aim of anonymizing the data and offering privacy protection. This thesis deals with databases containing semantically enriched trajectories that provide more information and remain stored in-house to the organization that created and recorded them. Based on studies and researches on simple trajectories, we develop a query auditing mechanism to avoid privacy violations, but at the same time suffer the least possible alteration in the data that must return as an answer to any such query. We are implementing a new algorithm, called Zoom_Out, which can be used to modify the original query when we are not allowed to give an answer for security reasons. The ultimate goal of the algorithm is to convert the original query to the closest possible query that cannot be compromised by privacy. In this way the mechanism becomes more flexible and user-friendly and enhances its functionality.