Offline hotspot analysis over road network trajectories
View/
Keywords
Hotspot detection ; Apache Spark ; Statistical significance ; Spatial networks ; Parallel processingAbstract
Given a dataset of vehicle trajectories on a road network the algorithms presented in this dissertation aim to discover statistically significant hotspots in said road network. This approach finds application in transportation engineering and the analysis of large collections of vehicle trajectories that can provide insights on the performance of road networks. Hotspots are identified using an adapted version of the Getis-Ord Gi* statistic, which has traditionally been used for the analysis of point data over a 2D grid. The statistic has been adapted to detect hotspots on a 3D spatio-temporal graph representation of the road network present in the input dataset. Two algorithms were developed; one that provides an exact solution to the problem and one that provides an approximate solution trading off accuracy for computational cost. The efficiency and scalability of the algorithms are tested experimentally on a vehicle trajectory dataset in Greece’s road network.
Postgraduate Studies Programme
Πληροφοριακά Συστήματα και ΥπηρεσίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνNumber of pages
51Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Attribution-NonCommercial-NoDerivatives 4.0 Διεθνές
Attribution-NonCommercial-NoDerivatives 4.0 Διεθνές