Back to search

Aristotle University Of Thessaloniki (AUTH)

User behavior analysis on social web with knowledge discovery techniques

Abstract

dc:description

The emergence of social media platforms changed drastically the way that people communicate. Due to the always-on always-connected devices, social interactions have migrated online, and social media has become an integral part of people's everyday life. The preoccupation with social media has resulted to an overwhelming sentiment expression. Users express opinions on various topics, while also emotional responses are transferred with diverse intensity. At the same time, however, new instantiations of negative interactions have arisen, including abusive behavior among online users. As new generations engage more with social media, abusive incidents are alas increasing, reaching epidemic levels. Since today's hyper-connected society allows human expression to occur anytime, anywhere, and instantly, the study of online human behavior will result to valuable knowledge discovery. Such discovered knowledge can empower new web services and applications with easily interpretable and comprehensible conclusions for the end users. Meanwhile, efficient methods should be developed so that the early detection and prevention of abusive behaviors become possible. This thesis addresses three problems: (i) sentiment detection from online text resources, (ii) discovery of abusive behaviors in social media, and (iii) empowering services with social, spatial, and temporal knowledge. Methods for data understanding, modeling, and knowledge extraction are proposed and go beyond the state of the art. As a source of information social media are mainly considered, spanning from microblogs and online social networks to content sharing platforms and location-based social networks. Both lexicon-based and machine learning methods are presented, either separately or under a hybrid scheme, to detect human opinions and emotions. To cover a wide human emotional spectrum, basic emotions, social ones, and those that characterize general affective states are analyzed. Moreover, in order to better conceptualize the structure of human emotion and the prevailing differences between the online and offline social interactions, an empirical examination of the emotion experience in computer mediated (online) and face-to-face (offline) social interactions takes place. A thorough study on diverse communities is presented which analyzes the properties of abusive users, the content they post, and how they differ from legitimate social media users. A robust methodology for extracting text, user, and network based attributes is proposed, studying the properties of bullies and aggressors, and what features distinguish them from regular users. In addition to measuring and modeling human sentiment expression and abusive behavior on the social web, the relationships that exist between social media users' activity and the real-world events are considered. At the same time, new services and applications are designed and developed that permit both the validation of the aforementioned methods and the imprinting of the discovered knowledge in an attractive and comprehensive way for the end-users. Overall, this thesis improves understanding of human sentiment expression in social media and reveals online abusive users and behaviors. New services and applications are built to connect the users' activity in social media with real-world events by considering social, spatial, and time-related characteristics.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2017

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Chatzakou, Despoina
  • Χατζάκου, Δέσποινα

Subjects

dc:subject × 15

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/42319

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Chatzakou, Despoina; Χατζάκου, Δέσποινα. User behavior analysis on social web with knowledge discovery techniques. Aristotle University Of Thessaloniki (AUTH), 2017. http://hdl.handle.net/10442/hedi/42319