Aristotle University Of Thessaloniki (AUTH)
Επεξεργασία ερωτημάτων ομοιότητας και κυριαρχίας σε πολυδιάστατους χώρους
Abstract
dc:descriptionThis thesis studies the processing of similarity and dominating queries between objects in spatial databases, in spatial networks and generally in multidimensional spaces. It presents the design, the analysis and the study of: (i) new similarity measures between trajectories of moving objects in space-time, with efficient processing algorithms and efficient indexing schemes, (ii) new selectivity estimation methods in spatial and general networks, (iii) a new efficient and progressive algorithm for graph node clustering via transitive node similarity, (iv) new efficient and progressive algorithms for processing of subspace top-k dominating queries in multidimensional datasets, which can be applied also in distributed environments, (v) new efficient and progressive algorithms of top-k dominating queries in metric spaces, and (vi) new estimation methods for the maximum domination value and the skyline cardinality in multidimensional datasets. The proposed algorithms and methods, for the processing of similarity and dominating queries, were designed for demanding applications with the following criteria: (i) to minimize the overall processing cost (CPU and I/O cost), (ii) to minimize the total required space in main memory for the query processing, (iii) to maximize the accuracy and the efficiency of the returned results, (iv) to minimize their complexity, (v) to provide the results in a progressive manner (as soon as the first results are available they can be returned to the user, while more answers are being prepared). The study for the processing of similarity and dominating queries has been completed with the extensive experimental evaluation and comparison of all new proposed methods and algorithms. Moreover, all the proposed methods and algorithms have been implemented in C++, and exhaustive experiments were performed for their comparison using real and synthetic datasets. The experimental results show that the proposed methods can be used efficiently in similarity and dominance applications as well as in demanding environments that manage these applications. In addition, the proposed methods have significant applications in different domains like: spatial databases, general databases, information retrieval, knowledge discovery, data mining, geographical information systems, road network applications, prediction systems, web applications, search engines, distributed and parallel environments, social networks, web communities discovery.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2011
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tiakas, Eleftherios
- Τιάκας, Ελευθέριος
Subjects
dc:subject × 18- Επεξεργασία ερωτήσεων
- Ομοιότητα
- Κυριαρχία
- Πολυδιάστατοι χώροι
- Χωρικά δίκτυα
- Χωρικές βάσεις δεδομένων
- Μετρικοί χώροι
- Query processing
- Similarity
- Domination
- Multidimensional spaces
- Spatial networks
- Spatial databases
- Metric spaces
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/25009