Aristotle University Of Thessaloniki (AUTH)
Τεχνικές εξόρυξης γνώσης από δεδομένα κοινωνικού ιστού
Abstract
dc:descriptionThe effect of digital technology in our everyday lives in the last decades has been profound. From the introduction of personal computers in the 80’s, to the emergence of the world wide web in the 90’s, to the widespread usage of the social web and web 2.0 technologies this decade, our lives have become strongly intertwined with the utilization as well as the production of shared data. Millions of users worldwide are posting every day to social networking sites, and are also tagging and annotating digital content. This huge amount of rich data which is strongly related to users, time, and space, could potentially become a highly invaluable source for extracting patterns and transforming them into knowledge towards responding to and solving a wide variety of modern-day problems. However, for such data to become useful, one cannot directly apply existing data mining techniques without special consideration of the peculiarities of the continuously evolving human-generated data sets of social networking sites, as well as of the target knowledge that would be useful to be derived from such data. Thus, in this thesis, we are presenting a number of techniques addressing exactly this need. These are comprised of four categories addressing communities, time, context and semantics. In more detail, first, algorithms are presented for clustering across the tag space, across combinations of tag and visual space, and across multiple social media-derived feature spaces, enabling community detection. Second, techniques for temporal analysis, which enable trend & event detection. Third, techniques for studying the sample size adequacy towards context extraction based on co-occurrence in social media, which can be used for enhancing real-world recognition problems. And fourth, techniques for synergism between the social and semantic web. All the above are introduced, discussed, and most importantly applied to real-world data, demonstrating their effectiveness. The toolchest provided by the above techniques, helps bridge the gap between massive raw social-network data and generic data mining techniques, on the one hand, and highly usable and useful high-level patterns and knowledge, on the other. Furthermore, multiple avenues for future research are presented and explored.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2013
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Giannakidou, Eirini
- Γιαννακίδου, Ειρήνη
Subjects
dc:subject × 16- Κοινωνικός ιστός
- Εξόρυξη δεδομένων
- Ομαδοποίηση
- Ανοιχτά δεδομένα
- Social web
- Data mining
- Clustering
- Open data
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Natural Sciences
- Computer and Information Sciences
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/38637