National and Kapodistrian University of Athens
Τεχνικές επεξεργασίας κατανεμημένων γράφων και ρευμάτων γράφων
Abstract
dc:descriptionBeneath most complex systems playing a vital role in our daily lives lie intricate networks. Such real-world networks are routinely represented using graphs. The volume of graph data produced in today’s interlinked world allows for realizing numerous fascinating applications but also posses important challenges. Consider for example the friendship graph of a social networking site and the findings we can come up with when executing network algorithms, such as community detection, on this graph. However, the volume that real-world networks reach oftentimes makes even the execution of fundamental graph algorithms infeasible when following traditional techniques.In this thesis we focus on two directions that allow for handling large scale networks, namely distributed graph processing, and streaming graph algorithms. In this context, we first provide contributions with regard to memory usage of distributed graph processing systems by extending the available structures of a contemporary such system with memory-optimized representations. Then, we focus on the task of community detection and propose i) a local algorithm that reveals the community structure of a vertex and easily facilitates distributed execution and ii) a streaming algorithm that greatly outperforms non-streaming state-of-the-art approaches with respect to both execution time and memory usage. In addition, we propose a streaming sampling technique that allows for capturing the interesting part of an unmanageable volume of data produced by social activity. Finally, we exploit the available data of a popular social networking site to empirically investigate a well-studied opinion formation model, using a distributed algorithm.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Liakos, Panagiotis
- Λιάκος, Παναγιώτης
Subjects
dc:subject × 14- Κατανεμημένη επεξεργασία γράφων
- Ρεύματα γράφων
- Συμπίεση γράφων
- Ανίχνευση κοινοτήτων
- Διαμόρφωση γνώμης
- Distributed graph processing
- Streaming graphs
- Graph compression
- Community detection
- Opinion formation
- 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/44433