Minimal spatio-temporal database repairs

Tobias Emrich, Hans-Peter Kriegel, Markus Mauder, Matthias Renz, Goce Trajcevski, Andreas Züfle
2013 Proceedings of the 21st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems - SIGSPATIAL'13  
This work addresses the problem of efficient detection and fixing of inconsistencies in Spatio-Temporal Databases. In contrast to traditional database settings where integrity constraints pertain to explicitly stored values (and, possibly, values defined via views and aggregates), we observe that spatio-temporal data may exhibit a specific types of violations to constraints that cannot be explicitly tied with the stored values. The main reason is that spatio-temporal phenomena are continuous,
more » ... t their database representations are discrete. Thus, the constraints are semantic in nature, as opposed to being firmly dependent on the actual stored data. We give a general definition of semantic constraints of a trajectory database, and define rules to repair violations of these constraints. We aim at minimizing the changes needed for repairing violations of such semantic constraints, in order do minimize the distortion of the state of the database induced by the repairs. Towards this, we define dissimilarity between the initial database and its repaired state and, to minimize this dissimilarity, we propose several rules of space-and time-distortion, which shift inconsistent observations in space and time to remove inconsistencies. Our evaluation shows that these simple rules often run into local minima, and thus may not be able to repair a database. To remedy this problem, we propose a hybrid approach which may choose between possible space and time distortions. We show that a greedy approach which always chooses the locally best repair may still run into local minima and propose a Simulated-Annealing approach, which combines greedy and random repairs to avoid these local minima.
doi:10.1145/2525314.2525468 dblp:conf/gis/EmrichKMRTZ13 fatcat:nvqlxdyllzhxrm6ejmucypi3va