Aristotle University Of Thessaloniki (AUTH)
Dominance-based knowledge discovery from large-scale attributed data
Abstract
dc:descriptionThis thesis establishes dominance relationships as a unifying and flexible mechanism for analyzing multi-attribute data in graphs and high-dimensional spaces, bypassing the need for rigid metric assumptions. It demonstrates this versatility across three core tasks, beginning with a skyline-based dissimilarity approach for high-dimensional image descriptors that captures intrinsic variation and utilizes hashing-based indexing for scalability. Second, it introduces DomiCom, a dimension-agnostic framework that merges exact or approximate domination scores with k-core decomposition to detect cohesive, attribute-significant, and optionally overlapping top-k dominant communities in attributed networks. Finally, it extends this reasoning to local community search, employing both hop-based and random-walk strategies guided by domination scores and maximum k-cores to build high-quality communities around user-specified query nodes. Supported by algorithmic formulations, theoretical analyses, and extensive experiments on real datasets, the research proves that dominance is a robust foundation for multi-criteria reasoning and new algorithm design in complex data environments.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Georgiadis, Nikolaos
- Γεωργιάδης, Νικόλαος
Subjects
dc:subject × 14- Σχέσεις κυριαρχίας
- Ανάλυση ανομοιότητας εικόνων
- Κυρίαρχες κοινότητες
- Εντοπισμός τοπικών κοινοτήτων
- Dominance relationships
- Image dissimilarity analysis
- Dominant communities
- Local community detection
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Επιστήμη ηλεκτρονικών υπολογιστών
- Computer and Information Sciences
- Natural Sciences
- Computer Science
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/61229